While Multimodal Large Language Models (MLLMs) offer strong perception and reasoning capabilities for image-text input, Visual Question Answering (VQA) focusing on small image details still remains a challenge. Although visual cropping techniques seem promising, recent approaches have several limitations: the need for task-specific fine-tuning, low efficiency due to uninformed exhaustive search, or incompatibility with efficient attention implementations. We address these shortcomings by proposing a training-free visual cropping method, dubbed FOCUS, that leverages MLLM-internal representations to guide the search for the most relevant image region. This is accomplished in four steps: first, we identify the target object(s) in the VQA prompt; second, we compute an object relevance map using the key-value (KV) cache; third, we propose and rank relevant image regions based on the map; and finally, we perform the fine-grained VQA task using the top-ranked region. As a result of this informed search strategy, FOCUS achieves strong performance across four fine-grained VQA datasets and three types of MLLMs. It outperforms three popular visual cropping methods in both accuracy and efficiency, and matches the best-performing baseline, ZoomEye, while requiring 3 – 6.5 × less compute.

Recent adversarial attacks with real world applications are capable of deceiving deep neural networks (DNN), which often appear as printed stickers applied to objects in physical world. Though achieving high success rate in lab tests and limited field tests, such attacks have not been tested on multiple DNN architectures with a standard setup to unveil the common robustness and weakness points of both the DNNs and the attacks. Furthermore, realistic looking stickers applied by normal people as acts of vandalism are not studied to discover their potential risks as well the risk of optimizing the location of such realistic stickers to achieve the maximum performance drop. In this paper, (a) we study the case of realistic looking sticker application effects on traffic sign detectors performance; (b) we use traffic sign image classification as our use case and train and attack 11 of the modern architectures for our analysis; (c) by considering different factors like brightness, blurriness and contrast of the train images in our sticker application procedure, we show that simple image processing techniques can help realistic looking stickers fit into their background to mimic real world tests; (d) by performing structured synthetic and real-world evaluations, we study the difference of various traffic sign classes in terms of their crucial distinctive features among the tested DNNs.

Despite recent advancements, deep neural networks are not robust against adversarial perturbations. Many of the proposed adversarial defense approaches use computationally expensive training mechanisms that do not scale to complex real-world tasks such as semantic segmentation, and offer only marginal improvements. In addition, fundamental questions on the nature of adversarial perturbations and their relation to the network architecture are largely understudied. In this work, we study the adversarial problem from a frequency domain perspective. More specifically, we analyze discrete Fourier transform (DFT) spectra of several adversarial images and report two major findings: First, there exists a strong connection between a model architecture and the nature of adversarial perturbations that can be observed and addressed in the frequency domain. Second, the observed frequency patterns are largely image- and attack-type independent, which is important for the practical impact of any defense making use of such patterns. Motivated by these findings, we additionally propose an adversarial defense method based on the well-known Wiener filters that captures and suppresses adversarial frequencies in a data-driven manner. Our proposed method not only generalizes across unseen attacks but also excels five existing state-of-the-art methods across two models in a variety of attack settings.

Instance segmentation with neural networks is an essential task in environment perception. In many works, it has been observed that neural networks can predict false positive instances with high confidence values and true positives with low ones. Thus, it is important to accurately model the uncertainties of neural networks in order to prevent safety issues and foster interpretability. In applications such as automated driving, the reliability of neural networks is of highest interest. In this paper, we present a time-dynamic approach to model uncertainties of instance segmentation networks and apply this to the detection of false positives as well as the estimation of prediction quality. The availability of image sequences in online applications allows for tracking instances over multiple frames. Based on an instances history of shape and uncertainty information, we construct temporal instance-wise aggregated metrics. The latter are used as input to post-processing models that estimate the prediction quality in terms of instance-wise intersection over union. The proposed method only requires a readily trained neural network (that may operate on single frames) and video sequence input. In our experiments, we further demonstrate the use of the proposed method by replacing the traditional score value from object detection and thereby improving the overall performance of the instance segmentation network.

In recent years, deep learning methods have outperformed other methods in image recognition. This has fostered imagination of potential application of deep learning technology including safety relevant applications like the interpretation of medical images or autonomous driving. The passage from assistance of a human decision maker to ever more automated systems however increases the need to properly handle the failure modes of deep learning modules. In this contribution, we review a set of techniques for the self-monitoring of machine-learning algorithms based on uncertainty quantification. In particular, we apply this to the task of semantic segmentation, where the machine learning algorithm decomposes an image according to semantic categories. We discuss false positive and false negative error modes at instance-level and review techniques for the detection of such errors that have been recently proposed by the authors. We also give an outlook on future research directions.

Artificial intelligence for autonomous driving must meet strict requirements on safety and robustness, which motivates the thorough validation of learned models. However, current validation approaches mostly require ground truth data and are thus both cost-intensive and limited in their applicability. We propose to overcome these limitations by a model agnostic validation using a-priori knowledge from street maps. In particular, we show how to validate semantic segmentation masks and demonstrate the potential of our approach using OpenStreetMap. We introduce validation metrics that indicate false positive or negative road segments. Besides the validation approach, we present a method to correct the vehicle's GPS position so that a more accurate localization can be used for the street-map based validation. Lastly, we present quantitative results on the Cityscapes dataset indicating that our validation approach can indeed uncover errors in semantic segmentation masks.

In semantic segmentation datasets, classes of high importance are oftentimes underrepresented, e.g., humans in street scenes. Neural networks are usually trained to reduce the overall number of errors, attaching identical loss to errors of all kinds. However, this is not necessarily aligned with human intuition. For instance, an overlooked pedestrian seems more severe than an incorrectly detected one. One possible remedy is to deploy different decision rules by introducing class priors that assign more weight to underrepresented classes. While reducing the false negatives of the underrepresented class, at the same time this leads to a considerable increase of false positive indications. In this work, we combine decision rules with methods for false positive detection. Therefore, we fuse false negative detection with uncertainty based false positive meta classification. We present the efficiency of our method for the semantic segmentation of street scenes on the Cityscapes dataset based on predicted instances of the "human" class. In the latter we employ an advanced false positive detection method using uncertainty measures aggregated over instances. We, thereby, achieve improved trade-offs between false negative and false positive samples of the underrepresented classes.

We present a method that "meta" classifies whether segments predicted by a semantic segmentation neural network intersect with the ground truth. For this purpose, we employ measures of dispersion for predicted pixel-wise class probability distributions, like classification entropy, that yield heat maps of the input scene's size. We aggregate these dispersion measures segment-wise and derive metrics that are well correlated with the segment-wise IoU of prediction and ground truth. This procedure yields an almost plug and play post-processing tool to rate the prediction quality of semantic segmentation networks on segment level. This is especially relevant for monitoring neural networks in online applications like automated driving or medical imaging where reliability is of utmost importance. In our tests, we use publicly available state-of-the-art networks trained on the Cityscapes dataset and the BraTS2017 dataset and analyze the predictive power of different metrics as well as different sets of metrics. To this end, we compute logistic LASSO regression fits for the task of classifying IoU = 0 vs. IoU > 0 per segment and obtain AUROC values of up to 91.55%. We complement these tests with linear regression fits to predict the segment-wise IoU and obtain prediction standard deviations of down to 0.130 as well as R2values of up to 84.15%. We show that these results clearly outperform standard approaches.

Cooperative perception with many sensors involved greatly improves the performance of perceptual systems in autonomous vehicles. However, the increasing amount of sensor data leads to a bottleneck due to limited capacity of vehicle-to-X (V2X) communication channels. We leverage lossy learned image compression by means of an autoencoder with adversarial loss function to reduce the overall bitrate. Our key contribution is to focus image compression on regions of interest (ROIs) governed by a binary mask. A transmitter-sided semantic segmentation network extracts semantically important classes being the basis for the generation of a ROI. A second key contribution is that the mask is not transmitted as side information, only the quantized bottleneck data is transmitted. To train the network, we use a loss function operating only on the pixels in the ROI. We report peak-signal-to-noise ratio (PSNR) both in the entire image and only in the ROI, evaluating various fusion architectures and fusion operations involving input image and mask. Showing the high generalizability of our approach, we achieve consistent improvements in the ROI in all experiments on the Cityscapes dataset.

The high amount of sensors required for autonomous driving poses enormous challenges on the capacity of automotive bus systems. There is a need to understand tradeoffs between bitrate and perception performance. In this paper, we compare the image compression standards JPEG, JPEG2000, and WebP to a modern encoder/decoder image compression approach based on generative adversarial networks (GANs). We evaluate both the pure compression performance using typical metrics such as peak signal-to-noise ratio (PSNR), structural similarity (SSIM) and others, but also the performance of a subsequent perception function, namely a semantic segmentation (characterized by the mean intersection over union (mIoU) measure). Not surprisingly, for all investigated compression methods, a higher bitrate means better results in all investigated quality metrics. Interestingly, however, we show that the semantic segmentation mIoU of the GAN autoencoder in the highly relevant low-bitrate regime (at 0.0625 bit/pixel) is better by 3.9 % absolute than JPEG2000, although the latter still is considerably better in terms of PSNR (5.91dB difference). This effect can greatly be enlarged by training the semantic segmentation model with images originating from the decoder, so that the mIoU using the segmentation model trained by GAN reconstructions exceeds the use of the model trained with original images by almost 20 % absolute. We conclude that distributed perception in future autonomous driving will most probably not provide a solution to the automotive bus capacity bottleneck by using standard compression schemes such as JPEG2000, but requires modern coding approaches, with the GAN encoder/decoder method being a promising candidate.

At present, devices are establishing themselves on the mobile device market which permit personalization and expansion by means of applications. These applications are increasingly also expected in the vehicle. In contrast to mobile devices, an expandable system in the vehicle is subject to special requirements. These are discussed in this publication and existing approaches are evaluated with regard to the requirements. In this case, outplacing subsequent applications to a smartphone or web server is identified as the best solution. For user interface (UI) integration to the driver information system, final user interface descriptions are identified as a good compromise between complexity and reusability. An improved approach for UI transfer with final UI descriptions using HTML is presented. Compared to existing concepts, this achieves an improvement in the form of graphical quality, response behaviour, reusability and complexity. A prototype implementation makes it clear that a series production system can be subsequently expanded with applications that cannot be distinguished from the basic functions of the driver information system in terms of graphical quality and response behaviour.