Integrating language and images for Composed Image Retrieval (CIR) allows for a flexible representation of search intent, making it a focal point in multi-modal research. Traditional CIR methods train models on complex triplet datasets. In contrast, Zero-Shot Composed Image Retrieval (ZS-CIR) eliminates the need for constructing datasets and training models for each specific task, attracting significant attention from researchers. CIReVL is a training-free method that translates visual content into textual representations, employing large language models (LLMs) to perform robust text inference and visual language models (VLMs) to ensure multi-modal alignment during retrieval. Although this approach achieves high performance without training and offers strong interpretability, relying solely on single descriptive texts for retrieval has limitations in handling complex search demands. To address this issue, we propose a DualCIR framework, which utilizes an LLM to generate refined dual-directional descriptions-both positive and negative-alongside generating holistic target descriptions. We score retrieval data separately via these descriptions, merge the scores, perform ranking, and ultimately achieve retrieval. Extensive experiments across three datasets in the natural image and fashion domains show that our approach maintains low inference costs while noticeably enhancing retrieval effectiveness. Compared to existing training-free methods, our DualCIR achieves state-of-the-art performance and is even competitive with the training-based methods.

The size of spatial receptive fields, from the early 3$\times$3 convolutions in VGGNet to the recent 7$\times$7 convolutions in ConvNeXt, has always played a critical role in architecture design. In this paper, we propose a Mixture of Receptive Fields (MoRF) instead of using a single receptive field. MoRF contains the combinations of multiple receptive fields with different sizes, e.g., convolutions with different kernel sizes, which can be regarded as experts. Such an approach serves two functions: one is to select the appropriate receptive field according to the input, and the other is to expand the network capacity. Furthermore, we also introduce two types of routing mechanisms, hard routing and soft routing to automatically select the appropriate receptive field experts. In the inference stage, the selected receptive field experts are merged via re-parameterization to maintain a similar inference speed compared to the single receptive field. To demonstrate the effectiveness of MoRF, we integrate the MoRF concept into multiple architectures, e.g., ResNet and ConvNeXt. Extensive experiments show that our approach outperforms the baselines in image classification, object detection, and segmentation tasks without significantly increasing the inference time.

Holistic scene understanding includes semantic segmentation, surface normal estimation, object boundary detection, depth estimation, etc. The key aspect of this problem is to learn representation effectively, as each subtask builds upon not only correlated but also distinct attributes. Inspired by visual-prompt tuning, we propose a Task-Specific Prompts Transformer, dubbed TSP-Transformer, for holistic scene understanding. It features a vanilla transformer in the early stage and tasks-specific prompts transformer encoder in the lateral stage, where tasks-specific prompts are augmented. By doing so, the transformer layer learns the generic information from the shared parts and is endowed with task-specific capacity. First, the tasks-specific prompts serve as induced priors for each task effectively. Moreover, the task-specific prompts can be seen as switches to favor task-specific representation learning for different tasks. Extensive experiments on NYUD-v2 and PASCAL-Context show that our method achieves state-of-the-art performance, validating the effectiveness of our method for holistic scene understanding. We also provide our code in the following link1.

Current audiovisual separation methods share a standard architecture design where an audio encoder-decoder network is fused with visual encoding features at the encoder bottleneck. This design confounds the learning of multimodal feature encoding with robust sound decoding for audio separation. To generalize to a new instrument, one must finetune the entire visual and audio network for all musical instruments. We re-formulate the visual-sound separation task and propose Instruments as Queries (iQuery) with a flexible query expansion mechanism. Our approach ensures cross-modal consistency and cross-instrument disentanglement. We utilize “visually named” queries to initiate the learning of audio queries and use cross-modal attention to remove potential sound source interference at the estimated waveforms. To generalize to a new instrument or event class, drawing inspiration from the text-prompt design, we insert additional queries as audio prompts while freezing the attention mechanism. Experimental results on three benchmarks demonstrate that our iQuery improves audiovisual sound source separation performance. Code is available at https://github.com/JiabenChen/iQuery.

Sequential video understanding, as an emerging video understanding task, has driven lots of researchers' attention because of its goal-oriented nature. This paper studies weakly supervised sequential video understanding where the accurate time-stamp level text-video alignment is not provided. We solve this task by borrowing ideas from CLIP. Specifically, we use a transformer to aggregate frame-level features for video representation and use a pre-trained text encoder to encode the texts corresponding to each action and the whole video, respectively. To model the correspondence between text and video, we propose a multiple granularity loss, where the video-paragraph contrastive loss enforces matching between the whole video and the complete script, and a fine-grained frame-sentence contrastive loss enforces the matching between each action and its description. As the frame-sentence correspondence is not available, we propose to use the fact that video actions happen sequentially in the temporal domain to generate pseudo frame-sentence correspondence and supervise the network training with the pseudo labels. Extensive experiments on video sequence verification and text-to-video matching show that our method outperforms baselines by a large margin, which validates the effectiveness of our proposed approach. Code is available at https://github.com/svip-lab/WeakSVR.

We propose a novel task for generating 3D dance movements that simultaneously incorporate both text and music modalities. Unlike existing works that generate dance movements using a single modality such as music, our goal is to produce richer dance movements guided by the instructive information provided by the text. However, the lack of paired motion data with both music and text modalities limits the ability to generate dance movements that integrate both. To alleviate this challenge, we propose to utilize a 3D human motion VQ-VAE to project the motions of the two datasets into a latent space consisting of quantized vectors, which effectively mix the motion tokens from the two datasets with different distributions for training. Additionally, we propose a cross-modal transformer to integrate text instructions into motion generation architecture for generating 3D dance movements without degrading the performance of music-conditioned dance generation. To better evaluate the quality of the generated motion, we introduce two novel metrics, namely Motion Prediction Distance (MPD) and Freezing Score (FS), to measure the coherence and freezing percentage of the generated motion. Extensive experiments show that our approach can generate realistic and coherent dance movements conditioned on both text and music while maintaining comparable performance with the two single modalities. Code is available at https://garfield-kh.github.io/TM2D/.

Text-to-motion generation is a formidable task, aiming to produce human motions that align with the input text while also adhering to human capabilities and physical laws. While there have been advancements in diffusion models, their application in discrete spaces remains underexplored. Current methods often overlook the varying significance of different motions, treating them uniformly. It is essential to recognize that not all motions hold the same relevance to a particular textual description. Some motions, being more salient and informative, should be given precedence during generation. In response, we introduce a Priority-Centric Motion Discrete Diffusion Model (M2DM), which utilizes a Transformer-based VQ-VAE to derive a concise, discrete motion representation, incorporating a global self-attention mechanism and a regularization term to counteract code collapse. We also present a motion discrete diffusion model that employs an innovative noise schedule, determined by the significance of each motion token within the entire motion sequence. This approach retains the most salient motions during the reverse diffusion process, leading to more semantically rich and varied motions. Additionally, we formulate two strategies to gauge the importance of motion tokens, drawing from both textual and visual indicators. Comprehensive experiments on the HumanML3D and KIT-ML datasets confirm that our model surpasses existing techniques in fidelity and diversity, particularly for intricate textual descriptions.

State-of-the-art deep neural networks are trained with large amounts (millions or even billions) of data. The expensive computation and memory costs make it difficult to train them on limited hardware resources, especially for recent popular large language models (LLM) and computer vision models (CV). Recent popular dataset distillation methods are thus developed, aiming to reduce the number of training samples via synthesizing small-scale datasets via gradient matching. However, as the gradient calculation is coupled with the specific network architecture, the synthesized dataset is biased and performs poorly when used for training unseen architectures. To address these limitations, we present dataset quantization (DQ), a new framework to compress large-scale datasets into small subsets which can be used for training any neural network architectures. Extensive experiments demonstrate that DQ is able to generate condensed small datasets for training unseen network architectures with state-of-the-art compression ratios for lossless model training. To the best of our knowledge, DQ is the first method that can successfully distill large-scale datasets such as ImageNet-1k with a state-of-the-art compression ratio. Notably, with 60% data from ImageNet and 20% data from Alpaca’s instruction tuning data, the models can be trained with negligible or no performance drop for both vision tasks (including classification, semantic segmentation, and object detection) as well as language tasks (including instruction tuning tasks such as BBH and DROP).

Dynamic vision sensors or event cameras provide rich complementary information for video frame interpolation. Existing state-of-the-art methods follow the paradigm of combining both synthesis-based and warping networks. However, few of those methods fully respect the intrinsic characteristics of events streams. Given that event cameras only encode intensity changes and polarity rather than color intensities, estimating optical flow from events is arguably more difficult than from RGB information. We therefore propose to incorporate RGB information in an event-guided optical flow refinement strategy. Moreover, in light of the quasi-continuous nature of the time signals provided by event cameras, we propose a divide-and-conquer strategy in which event-based intermediate frame synthesis happens incrementally in multiple simplified stages rather than in a single, long stage. Extensive experiments on both synthetic and real-world datasets show that these modifications lead to more reliable and realistic intermediate frame results than previous video frame interpolation methods. Our findings underline that a careful consideration of event characteristics such as high temporal density and elevated noise benefits interpolation accuracy.

Adapter-style efficient transfer learning (ETL) has shown excellent performance in the tuning of vision-language models (VLMs) under the low-data regime, where only a few additional parameters are introduced to excavate the task-specific knowledge based on the general and powerful representation of VLMs. However, most adapter-style works face two limitations: (i) modeling task-specific knowledge with a single modality only; and (ii) overlooking the exploitation of the inter-class relationships in downstream tasks, thereby leading to sub-optimal solutions. To mitigate that, we propose an effective adapter-style tuning strategy, dubbed GraphAdapter, which performs the textual adapter by explicitly modeling the dual-modality structure knowledge (i.e., the correlation of different semantics/classes in textual and visual modalities) with a dual knowledge graph. In particular, the dual knowledge graph is established with two sub-graphs, i.e., a textual knowledge sub-graph, and a visual knowledge sub-graph, where the nodes and edges represent the semantics/classes and their correlations in two modalities, respectively. This enables the textual feature of each prompt to leverage the task-specific structure knowledge from both textual and visual modalities, yielding a more effective classifier for downstream tasks. Extensive experimental results on 11 benchmark datasets reveal that our GraphAdapter significantly outperforms the previous adapter-based methods.

Counting repetitive actions are widely seen in human activities such as physical exercise. Existing methods focus on performing repetitive action counting in short videos, which is tough for dealing with longer videos in more realistic scenarios. In the data-driven era, the degradation of such generalization capability is mainly attributed to the lack of long video datasets. To complement this margin, we introduce a new large-scale repetitive action counting dataset covering a wide variety of video lengths, along with more realistic situations where action interruption or action inconsistencies occur in the video. Besides, we also provide a fine-grained annotation of the action cycles instead of just counting annotation along with a numerical value. Such a dataset contains 1,451 videos with about 20,000 annotations, which is more challenging. For repetitive action counting towards more realistic scenarios, we further propose encoding multi-scale temporal correlation with transformers that can take into account both performance and efficiency. Furthermore, with the help of fine-grained annotation of action cycles, we propose a density map regression-based method to predict the action period, which yields better performance with sufficient interpretability. Our proposed method outperforms state-of-the-art methods on all datasets and also achieves better performance on the unseen dataset without fine-tuning. The dataset and code are available11https://svip-lab.github.io/dataset/RepCount_dataset.html.

In this paper, we propose a novel sequence verification task that aims to distinguish positive video pairs performing the same action sequence from negative ones with step-level transformations but still conducting the same task. Such a challenging task resides in an open-set setting without prior action detection or segmentation that requires event-level or even frame-level annotations. To that end, we carefully reorganize two publicly available action-related datasets with step-procedure-task structure. To fully investigate the effectiveness of any method, we collect a scripted video dataset enumerating all kinds of step-level transformations in chemical experiments. Besides, a novel evaluation metric Weighted Distance Ratio is introduced to ensure equivalence for different step-level transformations during evaluation. In the end, a simple but effective baseline based on the transformer encoder with a novel sequence alignment loss is introduced to better characterize long-term dependency between steps, which outperforms other action recognition methods. Codes and data will be released1:

Existing fine-tuning methods either tune all parameters of the pre-trained model (full fine-tuning), which is not efficient, or only tune the last linear layer (linear probing), which suffers a significant accuracy drop compared to the full fine-tuning. In this paper, we propose a new parameter-efficient fine-tuning method termed as SSF, representing that researchers only need to Scale and Shift the deep Features extracted by a pre-trained model to catch up with the performance of full fine-tuning. In this way, SSF also surprisingly outperforms other parameter-efficient fine-tuning approaches even with a smaller number of tunable parameters. Furthermore, different from some existing parameter-efficient fine-tuning methods (e.g., Adapter or VPT) that introduce the extra parameters and computational cost in the training and inference stages, SSF only adds learnable parameters during the training stage, and these additional parameters can be merged into the original pre-trained model weights via re-parameterization in the inference phase. With the proposed SSF, our model obtains 2.46% (90.72% vs. 88.54%) and 11.48% (73.10% vs. 65.57%) performance improvement on FGVC and VTAB-1k in terms of Top-1 accuracy compared to the full fine-tuning but only fine-tuning about 0.3M parameters. We also conduct amounts of experiments in various model families (CNNs, Transformers, and MLPs) and datasets. Results on 26 image classification datasets in total and 3 robustness & out-of-distribution datasets show the effectiveness of SSF. Code is available at https://github.com/dongzelian/SSF.

To simultaneously estimate the number of heads and locate heads with bounding boxes, we resort to detection-based crowd counting by leveraging RGB-D data and design a dual-path guided detection network (DPDNet). Specifically, to improve the performance of detection-based approaches for dense/tiny heads, we propose a density map guided detection module, which leverages density map to improve the head/non-head classification in detection network where the density implies the probability of a pixel being a head, and a depth-adaptive kernel that considers the variances in head sizes is also introduced to generate high-fidelity density map for more robust density map regression. In order to prevent dense heads from being filtered out during post-processing, we utilize such a density map for post-processing of head detection and propose a density map guided NMS strategy. Meanwhile, to improve the ability of detecting small heads, we also propose a depth-guided detection module to generate a dynamic dilated convolution to extract features of heads of different scales, and a depth-aware anchor is further designed for better initialization of anchor sizes in the detection framework. Then we use the bounding boxes whose sizes are generated with depth to train our DPDNet. Considering that existing RGB-D datasets are too small and not suitable for performance evaluation of data-driven based approaches, we collect two large-scale RGB-D crowd counting datasets, which comprise a synthetic dataset and a real-world dataset, respectively. Since the depth value at long-distance positions cannot be obtained in the real-world dataset, we further propose a depth completion method with meta learning, which fully utilizes the synthetic depth data to complete the depth value at long-distance positions. Extensive experiments on our proposed two RGB-D datasets and the MICC RGB-D counting dataset show that our method achieves the best performance for RGB-D crowd counting and localization. Further, our method can be easily extended to RGB image based crowd counting and achieves comparable or even better performance on the RGB datasets for both head counting and localization.

Video Anomaly detection in videos refers to the identification of events that do not conform to expected behavior. However, almost all existing methods cast this problem as the minimization of reconstruction errors of training data including only normal events, which may lead to self-reconstruction and cannot guarantee a larger reconstruction error for an abnormal event. In this paper, we propose to formulate the video anomaly detection problem within a regime of video prediction. We advocate that not all video prediction networks are suitable for video anomaly detection. Then, we introduce two principles for the design of a video prediction network for video anomaly detection. Based on them, we elaborately design a video prediction network with appearance and motion constraints for video anomaly detection. Further, to promote the generalization of the prediction-based video anomaly detection for novel scenes, we carefully investigate the usage of a meta learning within our framework, where our model can be fast adapted to a new testing scene with only a few starting frames. Extensive experiments on both a toy dataset and three real datasets validate the effectiveness of our method in terms of robustness to the uncertainty in normal events and the sensitivity to abnormal events.

Locating and classifying clothes, usually referred to as clothing detection, is a fundamental task in fashion analysis. Motivated by the strong structural characteristics of clothes, we pursue a detection method enhanced by clothing keypoints, which is a compact and effective representation of structures. To incorporate the keypoint cues into clothing detection, we design a simple yet effective Keypoint-Guided clothing Detector, named KGDet. Such a detector can fully utilize information provided by keypoints with the following two aspects: i) integrating local features around keypoints to benefit both classification and regression; ii) generating accurate bounding boxes from keypoints. To effectively incorporate local features , two alternative modules are proposed. One is a multi-column keypoint-encoding-based feature aggregation module; the other is a keypoint-selection-based feature aggregation module. With either of the above modules as a bridge, a cascade strategy is introduced to refine detection performance progressively. Thanks to the keypoints, our KGDet obtains superior performance on the DeepFashion2 dataset and the FLD dataset with high efficiency.

An LBYL (‘Look Before You Leap’) Network is proposed for end-to-end trainable one-stage visual grounding. The idea behind LBYL-Net is intuitive and straightforward: we follow a language’s description to localize the target object based on its relative spatial relation to ‘Landmarks’, which is characterized by some spatial positional words and some descriptive words about the object. The core of our LBYL-Net is a landmark feature convolution module that transmits the visual features with the guidance of linguistic description along with different directions. Consequently, such a module encodes the relative spatial positional relations between the current object and its context. Then we combine the contextual information from the landmark feature convolution module with the target’s visual features for grounding. To make this landmark feature convolution light-weight, we introduce a dynamic programming algorithm (termed dynamic max pooling) with low complexity to extract the landmark feature. Thanks to the landmark feature convolution module, we mimic the human behavior of ‘Look Before You Leap’ to design an LBYL-Net, which takes full consideration of contextual information. Extensive experiments show our method’s effectiveness in four grounding datasets. Specifically, our LBYL-Net outperforms all state-of-the-art two-stage and one-stage methods on ReferitGame. On RefCOCO and RefCOCO+, Our LBYL-Net also achieves comparable results or even better results than existing one-stage methods. Code is available at https://github.com/svip-lab/LBYLNet.

To fully leverage the data captured from different scenes with different view angles while reducing the annotation cost, this paper studies a novel crowd counting setting, i.e. only using partial annotations in each image as training data. Inspired by the repetitive patterns in the annotated and unannotated regions as well as the ones between them, we design a network with three components to tackle those unannotated regions: i) in an Unannotated Regions Characterization (URC) module, we employ a memory bank to only store the annotated features, which could help the visual features extracted from these annotated regions flow to these unannotated regions; ii) For each image, Feature Distribution Consistency (FDC) regularizes the feature distributions of annotated head and unannotated head regions to be consistent; iii) a Cross-regressor Consistency Regularization (CCR) module is designed to learn the visual features of unannotated regions in a self-supervised style. The experimental results validate the effectiveness of our proposed model under the partial annotation setting for several datasets, such as ShanghaiTech, UCF-CC-50, UCF-QNRF, NWPU-Crowd and JHU-CROWD++. With only 10% annotated regions in each image, our proposed model achieves better performance than the recent methods and baselines under semi-supervised or active learning settings on all datasets. The code is https://github.com/svip-lab/CrwodCountingPAL.

This paper presents an anomaly detection method that is based on a sparse coding inspired Deep Neural Networks (DNN). Specifically, in light of the success of sparse coding based anomaly detection, we propose a Temporally-coherent Sparse Coding (TSC), where a temporally-coherent term is used to preserve the similarity between two similar frames. The optimization of sparse coefficients in TSC with the Sequential Iterative Soft-Thresholding Algorithm (SIATA) is equivalent to a special stacked Recurrent Neural Networks (sRNN) architecture. Further, to reduce the computational cost in alternatively updating the dictionary and sparse coefficients in TSC optimization and to alleviate hyperparameters selection in TSC, we stack one more layer on top of the TSC-inspired sRNN to reconstruct the inputs, and arrive at an sRNN-AE. We further improve sRNN-AE in the following aspects: i) rather than using a predefined similarity measurement between two frames, we propose to learn a data-dependent similarity measurement between neighboring frames in sRNN-AE to make it more suitable for anomaly detection; ii) to reduce computational costs in the inference stage, we reduce the depth of the sRNN in sRNN-AE and, consequently, our framework achieves real-time anomaly detection; iii) to improve computational efficiency, we conduct temporal pooling over the appearance features of several consecutive frames for summarizing information temporally, then we feed appearance features and temporally summarized features into a separate sRNN-AE for more robust anomaly detection. To facilitate anomaly detection evaluation, we also build a large-scale anomaly detection dataset which is even larger than the summation of all existing datasets for anomaly detection in terms of both the volume of data and the diversity of scenes. Extensive experiments on both a toy dataset under controlled settings and real datasets demonstrate that our method significantly outperforms existing methods, which validates the effectiveness of our sRNN-AE method for anomaly detection. Codes and data have been released at https://github.com/StevenLiuWen/sRNN_TSC_Anomaly_Detection.

This paper tackles RGBD based gaze estimation with Convolutional Neural Networks (CNNs). Specifically, we propose to decompose gaze point estimation into eyeball pose, head pose, and 3D eye position estimation. Compared with RGB image-based gaze tracking, having depth modality helps to facilitate head pose estimation and 3D eye position estimation. The captured depth image, however, usually contains noise and black holes which noticeably hamper gaze tracking. Thus we propose a CNN-based multi-task learning framework to simultaneously refine depth images and predict gaze points. We utilize a generator network for depth image generation with a Generative Neural Network (GAN), where the generator network is partially shared by both the gaze tracking network and GAN-based depth synthesizing. By optimizing the whole network simultaneously, depth image synthesis improves gaze point estimation and vice versa. Since the only existing RGBD dataset (EYEDIAP) is too small, we build a large-scale RGBD gaze tracking dataset for performance evaluation. As far as we know, it is the largest RGBD gaze dataset in terms of the number of participants. Comprehensive experiments demonstrate that our method outperforms existing methods by a large margin on both our dataset and the EYEDIAP dataset.

We propose a new learning paradigm, Local to Global Learning (LGL), for Deep Neural Networks (DNNs) to improve the performance of classification problems. The core of LGL is to learn a DNN model from fewer categories (local) to more categories (global) gradually within the entire training set. LGL is most related to the Self-Paced Learning (SPL) algorithm but its formulation is different from SPL. SPL trains its data from simple to complex, while LGL from local to global. In this paper, we incorporate the idea of LGL into the learning objective of DNNs and explain why LGL works better from an information-theoretic perspective. Experiments on the toy data, CIFAR-10, CIFAR-100, and ImageNet dataset show that LGL outperforms the baseline and SPL-based algorithms.

To simultaneously estimate head counts and localize heads with bounding boxes, a regression guided detection network (RDNet) is proposed for RGB-D crowd counting. Specifically, to improve the robustness of detection-based approaches for small/tiny heads, we leverage density map to improve the head/non-head classification in detection network where density map serves as the probability of a pixel being a head. A depth-adaptive kernel that considers the variances in head sizes is also introduced to generate high-fidelity density map for more robust density map regression. Further, a depth-aware anchor is designed for better initialization of anchor sizes in detection framework. Then we use the bounding boxes whose sizes are estimated with depth to train our RDNet. The existing RGB-D datasets are too small and not suitable for performance evaluation on data-driven based approaches, we collect a large-scale RGB-D crowd counting dataset. Experiments on both our RGB-D dataset and the MICC RGB-D counting dataset show that our method achieves the best performance for RGB-D crowd counting and localization. Further, our method can be readily extended to RGB image based crowd counting and achieves comparable performance on the ShanghaiTech Part\_B dataset for both counting and localization.

Single-image piece-wise planar 3D reconstruction aims to simultaneously segment plane instances and recover 3D plane parameters from an image. Most recent approaches leverage convolutional neural networks (CNNs) and achieve promising results. However, these methods are limited to detecting a fixed number of planes with certain learned order. To tackle this problem, we propose a novel two-stage method based on associative embedding, inspired by its recent success in instance segmentation. In the first stage, we train a CNN to map each pixel to an embedding space where pixels from the same plane instance have similar embeddings. Then, the plane instances are obtained by grouping the embedding vectors in planar regions via an efficient mean shift clustering algorithm. In the second stage, we estimate the parameter for each plane instance by considering both pixel-level and instance-level consistencies. With the proposed method, we are able to detect an arbitrary number of planes. Extensive experiments on public datasets validate the effectiveness and efficiency of our method. Furthermore, our method runs at 30 fps at the testing time, thus could facilitate many real-time applications such as visual SLAM and human-robot interaction. Code is available at https://github.com/svip-lab/PlanarReconstruction.

Gaze estimation, which aims to predict gaze points with given eye images, is an important task in computer vision because of its applications in human visual attention understanding. Many existing methods are based on a single camera, and most of them only focus on either the gaze point estimation or gaze direction estimation. In this paper, we propose a novel multitask method for the gaze point estimation using multiview cameras. Specifically, we analyze the close relationship between the gaze point estimation and gaze direction estimation, and we use a partially shared convolutional neural networks architecture to simultaneously estimate the gaze direction and gaze point. Furthermore, we also introduce a new multiview gaze tracking data set that consists of multiview eye images of different subjects. As far as we know, it is the largest multiview gaze tracking data set. Comprehensive experiments on our multiview gaze tracking data set and existing data sets demonstrate that our multiview multitask gaze point estimation solution consistently outperforms existing methods.

Anomaly detection in videos refers to the identification of events that do not conform to expected behavior. However, almost all existing methods tackle the problem by minimizing the reconstruction errors of training data, which cannot guarantee a larger reconstruction error for an abnormal event. In this paper, we propose to tackle the anomaly detection problem within a video prediction framework. To the best of our knowledge, this is the first work that leverages the difference between a predicted future frame and its ground truth to detect an abnormal event. To predict a future frame with higher quality for normal events, other than the commonly used appearance (spatial) constraints on intensity and gradient, we also introduce a motion (temporal) constraint in video prediction by enforcing the optical flow between predicted frames and ground truth frames to be consistent, and this is the first work that introduces a temporal constraint into the video prediction task. Such spatial and motion constraints facilitate the future frame prediction for normal events, and consequently facilitate to identify those abnormal events that do not conform the expectation. Extensive experiments on both a toy dataset and some publicly available datasets validate the effectiveness of our method in terms of robustness to the uncertainty in normal events and the sensitivity to abnormal events. All codes are released in https://github.com/StevenLiuWen/ano_pred_cvpr2018.