Multilayer neural networks trained with the back-propagation algorithm constitute the best example of a successful gradient based learning technique. Given an appropriate network architecture, gradient-based learning algorithms can be used to synthesize a complex decision surface that can classify high-dimensional patterns, such as handwritten characters, with minimal preprocessing. This paper reviews various methods applied to handwritten character recognition and compares them on a standard handwritten digit recognition task. Convolutional neural networks, which are specifically designed to deal with the variability of 2D shapes, are shown to outperform all other techniques. Real-life document recognition systems are composed of multiple modules including field extraction, segmentation recognition, and language modeling. A new learning paradigm, called graph transformer networks (GTN), allows such multimodule systems to be trained globally using gradient-based methods so as to minimize an overall performance measure. Two systems for online handwriting recognition are described. Experiments demonstrate the advantage of global training, and the flexibility of graph transformer networks. A graph transformer network for reading a bank cheque is also described. It uses convolutional neural network character recognizers combined with global training techniques to provide record accuracy on business and personal cheques. It is deployed commercially and reads several million cheques per day.

Gradient-based learning applied to document recognition

We propose a deep convolutional neural network architecture codenamed Inception that achieves the new state of the art for classification and detection in the ImageNet Large-Scale Visual Recognition Challenge 2014 (ILSVRC14). The main hallmark of this architecture is the improved utilization of the computing resources inside the network. By a carefully crafted design, we increased the depth and width of the network while keeping the computational budget constant. To optimize quality, the architectural decisions were based on the Hebbian principle and the intuition of multi-scale processing. One particular incarnation used in our submission for ILSVRC14 is called GoogLeNet, a 22 layers deep network, the quality of which is assessed in the context of classification and detection.

/pdf/going-deeper-with-convolutions-1yobw2o2ds.pdf

Going deeper with convolutions

There is, I think, something ethereal about i —the square root of minus one. I remember first hearing about it at school. It seemed an odd beast at that time—an intruder hovering on the edge of reality.

Usually familiarity dulls this sense of the bizarre, but in the case of i it was the reverse: over the years the sense of its surreal nature intensified. It seemed that it was impossible to write mathematics that described the real world in …

I and i

We study the question of feature sets for robust visual object recognition; adopting linear SVM based human detection as a test case. After reviewing existing edge and gradient based descriptors, we show experimentally that grids of histograms of oriented gradient (HOG) descriptors significantly outperform existing feature sets for human detection. We study the influence of each stage of the computation on performance, concluding that fine-scale gradients, fine orientation binning, relatively coarse spatial binning, and high-quality local contrast normalization in overlapping descriptor blocks are all important for good results. The new approach gives near-perfect separation on the original MIT pedestrian database, so we introduce a more challenging dataset containing over 1800 annotated human images with a large range of pose variations and backgrounds.

/pdf/histograms-of-oriented-gradients-for-human-detection-3nteinziuq.pdf

Histograms of oriented gradients for human detection

Object detection performance, as measured on the canonical PASCAL VOC dataset, has plateaued in the last few years. The best-performing methods are complex ensemble systems that typically combine multiple low-level image features with high-level context. In this paper, we propose a simple and scalable detection algorithm that improves mean average precision (mAP) by more than 30% relative to the previous best result on VOC 2012 -- achieving a mAP of 53.3%. Our approach combines two key insights: (1) one can apply high-capacity convolutional neural networks (CNNs) to bottom-up region proposals in order to localize and segment objects and (2) when labeled training data is scarce, supervised pre-training for an auxiliary task, followed by domain-specific fine-tuning, yields a significant performance boost. Since we combine region proposals with CNNs, we call our method R-CNN: Regions with CNN features. We also present experiments that provide insight into what the network learns, revealing a rich hierarchy of image features. Source code for the complete system is available at http://www.cs.berkeley.edu/~rbg/rcnn.

/pdf/rich-feature-hierarchies-for-accurate-object-detection-and-22fvjlggyg.pdf

Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation

Consider a stereotypical image-retrieval problem; a user submits a set of query images to a system and through repeated interactions during which the system presents its current choices and the user gives his/her preferences to them, the choices are narrowed to the image(s) that satisfies the user. The problem obviously must deal with image content, i.e., interpretation and preference. For this purpose, conventional so-called contentbased image retrieval (CBIR) approach uses image-processing and computer-vision techniques, and tries to understand the image content. Such attempts have produced good but limited success, mainly because image interpretation is a highly complicated perceptive process. We propose a new approach to this problem from a totally different angle. It attempts to exploit the human’s perceptual capabilities and certain common, if not identical, tendencies that must exist among people’s interpretation and preference of images. Instead of processing images, the system simply accumulates records of user feedback and recycles them in the form of collaborative filtering, just like a purchase recommendation system such as Amazo.com. To emphasize the point that it does not deal with image pixel information, we dub the approach by a term “content-free” image retrieval (CFIR). We discuss various issues of image retrieval, argue for the idea of CFIR, and present results of preliminary experiment. The results indicate that the performance of CFIR improves with the number of accumulated feedbacks, outperforming a basic but typical conventional CBIR system.

/pdf/user-powered-content-free-approach-to-image-retrieval-2ww8pbjenj.pdf

User-Powered "Content-Free" Approach to Image Retrieval

An overview of the development of a robot walker for use on the NASA Space Station is presented. The robot is designed to perform such tasks as inspection, transport of parts, and simple manipulation. The robot hardware and grippers, the real-time control hardware and software for precise, stable control of the highly flexible, three-dimensional robot, and a gravity compensation system are discussed

Development of a 5-DOF walking robot for Space Station application: overview

The need for robust self‐contained and low-latency vision systems is growing: high speed visual servoing and vision‐based human computer interface. Conventional vision systems can hardly meet this need because 1) the latency is incurred in a data transfer and computational bottlenecks, and 2) there is no top‐down feedback to adapt sensor performance for improved robustness. In this paper we present a tracking computational sensor — a VLSI implementation of a sensory attention. The tracking sensor focuses attention on a salient feature in its receptive field and maintains this attention in the world coordinates. Using both low ‐latency massive parallel processing and top‐down sensory adaptation, the sensor reliably tracks features of interest while it suppresses other irrelevant features that may interfere with the task at hand.

/pdf/sensory-attention-computational-sensor-paradigm-for-low-5d9flg65ur.pdf

Sensory Attention: Computational Sensor Paradigm for Low-Latency Adaptive Vision

Material transfer robots first appeared in the mid-1960s for use in traditional industrial applications. By the 1980s, robots found use in more demanding industrial applications such as welding, assembly, and inspection, with the help of vision and other sensors. It is estimated that 46,000 industrial robots have been installed in the U.S. japan has six to eight times as many robots

New technologies and applications in robotics

Objects in a real world image cannot have arbitrary appearance, sizes and locations due to geometric constraints in 3D space. Such a 3D geometric context plays an important role in resolving visual ambiguities and achieving coherent object detection. In this paper, we develop a RANSAC-CRF framework to detect objects that are geometrically coherent in the 3D world. Different from existing methods, we propose a novel generalized RANSAC algorithm to generate global 3D geometry hypotheses from local entities such that outlier suppression and noise reduction is achieved simultaneously. In addition, we evaluate those hypotheses using a CRF which considers both the compatibility of individual objects under global 3D geometric context and the compatibility between adjacent objects under local 3D geometric context. Experiment results show that our approach compares favorably with the state of the art.

/pdf/coherent-object-detection-with-3d-geometric-context-from-a-2wlvnpot1l.pdf

Takeo Kanade

Papers

User-Powered "Content-Free" Approach to Image Retrieval

Development of a 5-DOF walking robot for Space Station application: overview

Sensory Attention: Computational Sensor Paradigm for Low-Latency Adaptive Vision

New technologies and applications in robotics

Coherent Object Detection with 3D Geometric Context from a Single Image