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

抗原变异可使得多种致病微生物易于逃避宿主免疫应答。表达在感染红细胞表面的恶性疟原虫红细胞表面蛋白1（PfPMP1）与感染红细胞、内皮细胞、树突状细胞以及胎盘的单个或多个受体作用，在黏附及免疫逃避中起关键的作用。每个单倍体基因组var基因家族编码约60种成员，通过启动转录不同的var基因变异体为抗原变异提供了分子基础。

疟原虫var基因转换速率变化导致抗原变异[英]／Paul H, Robert P, Christodoulou Z, et al//Proc Natl Acad Sci U S A

Presents a theoretically very simple, yet efficient, multiresolution approach to gray-scale and rotation invariant texture classification based on local binary patterns and nonparametric discrimination of sample and prototype distributions. The method is based on recognizing that certain local binary patterns, termed "uniform," are fundamental properties of local image texture and their occurrence histogram is proven to be a very powerful texture feature. We derive a generalized gray-scale and rotation invariant operator presentation that allows for detecting the "uniform" patterns for any quantization of the angular space and for any spatial resolution and presents a method for combining multiple operators for multiresolution analysis. The proposed approach is very robust in terms of gray-scale variations since the operator is, by definition, invariant against any monotonic transformation of the gray scale. Another advantage is computational simplicity as the operator can be realized with a few operations in a small neighborhood and a lookup table. Experimental results demonstrate that good discrimination can be achieved with the occurrence statistics of simple rotation invariant local binary patterns.

/pdf/multiresolution-gray-scale-and-rotation-invariant-texture-1b1y2fta4g.pdf

Multiresolution gray-scale and rotation invariant texture classification with local binary patterns

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 compare R-CNN to OverFeat, a recently proposed sliding-window detector based on a similar CNN architecture. We find that R-CNN outperforms OverFeat by a large margin on the 200-class ILSVRC2013 detection dataset. Source code for the complete system is available at this http URL.

Rich feature hierarchies for accurate object detection and semantic segmentation

Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

In this paper, we propose an unsupervised approach for learning the three factors of the topology of a non-overlapping multi-camera network, which are nodes, links, and transition time distributions. It is a cross-correlation based method. Different from previous methods, the proposed method can deal with large amounts of data without considering the size of time window. The connectivity between nodes is estimated based on the N-neighbor accumulated cross-correlations, as well as the transition time distribution for each link. Furthermore, integrated with similarity cues, the proposed method can be extended into weighted cross-correlation models for better performance. Experimental results both on simulated and real-life datasets demonstrate the effectiveness of the proposed method.

/pdf/learning-the-three-factors-of-a-non-overlapping-multi-camera-1wnuj7dbg8.pdf

Learning the Three Factors of a Non-overlapping Multi-camera Network Topology

A novel algorithm is presented for determining the orientation of road vehicles in traffic scenes using video images. The algorithm requires no specific 3-D vehicle models and only uses local image gradient values. It may easily be implemented in real-time. Experimental results with a variety of vehicles in routine traffic scenes are included to demonstrate the effectiveness of the algorithm.

Model-independent recovery of object orientations

A method for detecting out sensitive audio based on breaking up motion colour of skin includes carrying out breaking up and border picking up for motion object in audio, detecting colour of skin for object broken up for deriving out exposing degree of skin comparing to motion object, making comprehensive evaluation for sensitivity of entire audio based on calculation of each frame. The method can be applied on Internet for identifying and filtering out information which is not civility.

Sensitive video frequency detection based on kinematic skin division

An anti-fraud method for printed matter features that the picture to be printed is processed by digital watermark technique and then printed by indistinctly visual ink Its advantage is high anti-fraud effect

Antifraud method for printed matter

Face verification is useful in a variety of applications. A face verification system is vulnerable not only to variations in ambient lighting, facial expression and facial pose, but also to the effect of small sample size during the training phase. In this paper, we propose an approach to face verification based on Radial Basis Function (RBF) networks and bagging. The technique seeks to offset the effect of using a small sample size during the training phase. The RBF networks are trained using all available positive samples of a subject and a few randomly selected negative samples. Bagging is then applied to the outputs of these RBF-based classifiers. Theoretical analysis and experimental results show the validity of the proposed approach.

/pdf/face-verification-based-on-bagging-rbf-networks-1ud2lbadw7.pdf

Tieniu Tan

Papers

Learning the Three Factors of a Non-overlapping Multi-camera Network Topology

Model-independent recovery of object orientations

Sensitive video frequency detection based on kinematic skin division

Antifraud method for printed matter

Face verification based on bagging RBF networks