In few-shot learning scenarios, the challenge is to generalize and perform well on new unseen examples when only very few labeled examples are available for each task. Model-agnostic meta-learning (MAML) has gained the popularity as one of the representative few-shot learning methods for its flexibility and applicability to diverse problems. However, MAML and its variants often resort to a simple loss function without any auxiliary loss function or regularization terms that can help achieve better generalization. The problem lies in that each application and task may require different auxiliary loss function, especially when tasks are diverse and distinct. Instead of attempting to hand-design an auxiliary loss function for each application and task, we introduce a new meta-learning framework with a loss function that adapts to each task. Our proposed framework, named Meta-Learning with Task-Adaptive Loss Function (MeTAL), demonstrates the effectiveness and the flexibility across various domains, such as few-shot classification and few-shot regression.

Scene text recognition is a challenging problem due to the wide variances in contents, styles, orientations, and image quality of text instances in natural scene images. To learn the intrinsic representation of scene texts, a novel multi-element attention (MEA) mechanism is proposed to exploit geometric structures from local to global levels in feature maps extracted from a scene text image. The MEA mechanism is a generalized form of self-attention technique. The elements in feature maps are taken as the nodes of an undirected graph, and three kinds of adjacency matrices are designed to aggregate information at local, neighborhood and global levels before calculating the attention weights. A multi-element attention network (MEAN) is implemented, which includes a CNN for feature extraction, an encoder with MEA mechanism and a decoder for predicting text codes. Orientational positional encoding is added to feature maps output by the CNN, and a feature vector sequence transformed from the feature maps is used as the input of the encoder. Experimental results show that MEAN has achieved state-of-the-art or competitive performance on seven public English scene text datasets (IIITSk, SVT, IC03, IC13, IC15, SVTP, and CUTE). Further experiments have been conducted on a selected subset of the RCTW Chinese scene text dataset, demonstrating that MEAN can handle horizontal, vertical, and irregular scene text samples.

We propose an image based end-to-end learning framework that helps lane-change decisions for human drivers and autonomous vehicles. The proposed system, Safe Lane-Change Aid Network (SLCAN), trains a deep convolutional neural network to classify the status of adjacent lanes from rear view images acquired by cameras mounted on both sides of the vehicle. Rather than depending on any explicit object detection or tracking scheme, SLCAN reads the whole input image and directly decides whether initiation of the lane-change at the moment is safe or not. We collected and annotated 77,273 rear side view images to train and test SLCAN. Experimental results show that the proposed framework achieves 96.98% classification accuracy although the test images are from unseen roadways. We also visualize the saliency map to understand which part of image SLCAN looks at for correct decisions.

Over the last couple of years, face recognition researchers have been developing new techniques. These developments are being fueled by advances in computer vision techniques, computer design, sensor design, and interest in fielding face recognition systems. Such advances hold the promise of reducing the error rate in face recognition systems by an order of magnitude over Face Recognition Vendor Test (FRVT) 2002 results. The face recognition grand challenge (FRGC) is designed to achieve this performance goal by presenting to researchers a six-experiment challenge problem along with data corpus of 50,000 images. The data consists of 3D scans and high resolution still imagery taken under controlled and uncontrolled conditions. This paper describes the challenge problem, data corpus, and presents baseline performance and preliminary results on natural statistics of facial imagery.

Previous performance evaluation of range image segmentation algorithms has depended on manual tuning of algorithm parameters, and has lacked a basis for a test of the significance of differences between algorithms. We present an automated framework for evaluating the performance of range image segmentation algorithms. Automated tuning of algorithm parameters in this framework results in performance as good as that previously obtained with careful manual tuning by the algorithm developers. Use of multiple training and test sets of images provides the basis for a test of the significance of performance differences between algorithms. The framework implementation includes range images, ground truth overlays, program source code, and shell scripts. This framework should a) make it possible to objectively and reliably compare the performance of range image segmentation algorithms; b) allow informed experimental feedback for the design of improved segmentation algorithms. The framework is demonstrated using range images, but in principle it could be used to evaluate region segmentation algorithms for any type of image.

We have developed an automated framework for performance evaluation of curved-surface range image segmentation algorithms. Enhancements over our previous work include automated training of parameter values, correcting the artifact problem in K/sup 2/T scanner images, and acquisition of images of the same scenes from different range scanners. The image dataset includes planar, spherical, cylindrical, conical, and toroidal surfaces. We have evaluated the automated parameter tuning technique and found that it compares favorably with manual parameter tuning. We present initial results from comparing curved-surface segmenters by Besl and Jain (1988) and by Jiang and Bunke (1998).

We have developed an automated framework for objectively evaluating the performance of region segmentation algorithms. This framework is demonstrated with range image data sets, but is applicable to any type of imagery. Parameters of the segmentation algorithm are tuned using training images. Images and source code for the training process care publicly available. The trained parameters are then used to evaluate the algorithm on a (sequestered) test set. The primary performance metric is the average number of correctly segmented regions. Statistical tests are used to determine the significance of performance improvement over a baseline algorithm.