Gaurav Sharma

Discriminative spatial saliency for image classification

In many visual classification tasks the spatial distribution of discriminative information is (i) non uniform e.g. person `reading can be distinguished from `taking a photo based on the area around the arms i.e. ignoring the legs and (ii) has intra class variations e.g. different readers may hold the books differently. Motivated by these observations, we propose to learn the discriminative spatial saliency of images while simultaneously learning a max margin classifier for a given visual classification task. Using the saliency maps to weight the corresponding visual features improves the discriminative power of the image representation. We treat the saliency maps as latent variables and allow them to adapt to the image content to maximize the classification score, while regularizing the change in the saliency maps. Our experimental results on three challenging datasets, for (i) human action classification, (ii) fine grained classification and (iii) scene classification, demonstrate the effectiveness and wide applicability of the method.

Local higher-order statistics (LHS) for texture categorization and facial analysis

This paper proposes a new image representation for texture categorization and facial analysis, relying on the use of higher-order local differential statistics as features. In contrast with models based on the global structure of textures and faces, it has been shown recently that small local pixel pattern distributions can be highly discriminative. Motivated by such works, the proposed model employs higher-order statistics of local non-binarized pixel patterns for the image description. Hence, in addition to being remarkably simple, it requires neither any user specified quantization of the space (of pixel patterns) nor any heuristics for discarding low occupancy volumes of the space. This leads to a more expressive representation which, when combined with discriminative SVM classifier, consistently achieves state-of-the-art performance on challenging texture and facial analysis datasets outperforming contemporary methods (with similar powerful classifiers).

Efficient Skin Region Segmentation Using Low Complexity Fuzzy Decision Tree Model

We propose an efficient skin region segmentation methodology using low complexity fuzzy decision tree constructed over B, G, R colour space. Skin and nonskin training dataset has been generated by using various skin textures obtained from face images of diversity of age, gender, and race people and nonskin pixels obtained from arbitrary thousands of random sampling of nonskin textures. Compact fuzzy model with very few numbers of rules allow to raster scan consumer photographs and classify each pixel as skin or nonskin for various face and human detection applications for embedded platforms.

A Novel Approach for Efficient SVM Classification with Histogram Intersection Kernel

The kernel trick - commonly used in machine learning and computer vision - enables learning of non-linear decision functions without having to explicitly map the original data to a high dimensional space. However, at test time, it requires evaluating the kernel with each one of the support vectors, which is time consuming. In this paper, we propose a novel approach for learning non-linear SVM corresponding to the histogram intersection kernel without using the kernel trick. We formulate the exact non-linear problem in the original space and show how to perform classification directly in this space. The learnt classifier incorporates non-linearity while maintaining O(d) testing complexity (for d-dimensional input space), compared to O(d Nsv) when using the kernel trick. We show that the SVM problem with histogram intersection kernel is quasi-convex in input space and outline an iterative algorithm to solve it. The proposed approach has been validated in experiments where it is compared with other linear SVM-based methods, showing that the proposed method achieves similar or better performance at lower computational and memory costs.

Adaptive Digital Makeup

A gender and skin color ethnicity based automatic digital makeup system is presented. An automatic face makeup system which applies example based digital makeup based on skin ethnicity color and gender type. One major advantage of the system is that the makeup is based on the skin color and gender type, which is very necessary for an effective makeup. Another strong advantage is that it applies automatic makeup without requiring any user input.

Curvature feature distribution based classification of Indian scripts from document images

We present a framework for classification of text document images based on their script. We deal with the domain of Indian scripts which has high inter script similarities. Indian scripts have characteristic curvature distributions which help in visual discrimination of scripts. We use edge direction based features to capture the distribution of curvature. We also use a recently proposed feature selection algorithm to obtain the most discriminating curvature features. We form hierarchy (automatically) based on statistical distances between the script models. Hierarchy allows us to group similar scripts at one level and then focus on the classification between the similar scripts at the next level leading to improvement in accuracy. We show experiments and results on a large set of about 3400 images.

Distributed calibration of pan-tilt camera network using multi-layered belief propagation

In this paper, we present a technique for distributed self-calibration of pan-tilt camera network using multi-layered belief propagation. Our goal is to obtain globally consistent estimates of the camera parameters for each camera with respect to a global world coordinate system. The network configuration changes with time as the cameras can pan and tilt. We also give a distributed algorithm for automatically finding which cameras have overlapping views at a certain point in time. We argue that using belief propagation it is sufficient to have correspondences between three cameras at a time for calibrating a larger set of (static) cameras with overlapping views. Our method gives an accurate and globally consistent estimate of the camera parameters of each camera in the network.

Bag-of-features kernel eigen spaces for classification

We present a classifier unifying local features based representation and subspace based learning. We also propose a novel method to merge kernel eigen spaces (KES) in feature space. Subspace methods have traditionally been used with the full appearance of the image. Recently local features based bag-of-features (BoF) representation has performed impressively on classification tasks. We use KES with BoF vectors to construct class specific subspaces and use the distance of a query vector from the database KESs as the classification criteria. The use of local features makes our approach invariant to illumination, rotation, scale, small affine transformation and partial occlusions. The system allows hierarchy by merging the KES in the feature space. The classifier performs competitively on the challenging Caltech-101 dataset under normal and simulated occlusion conditions. We show hierarchy on a dataset of videos collected over the internet.

Categorization and Reorientation of Images Based on Low Level Features

A hierarchical system to perform automatic categorization and reorientation of images using content analysis is pre-sented. The proposed system first categorizes images to some a priori defined categories using rotation invariant features. At the second stage, it detects their correct orientation out of {0o, 90o, 180o, and 270o} using category specific model. The system has been specially designed for embedded devices applications using only low level color and edge features. Machine learning algorithms optimized to suit the embedded implementation like support vector machines (SVMs) and scalable boosting have been used to develop classifiers for categorization and orientation detection. Results are presented on a collection of about 7000 consumer images collected from open resources. The proposed system finds it applications to various digital media products and brings pattern recognition solutions to the consumer electronics domain.

Hierarchical System for Content Based Categorization and Orientation of Consumer Images

A hierarchical framework to perform automatic categorization and reorientation of consumer images based on their content is presented. Sometimes the consumer rotates the camera while taking the photographs but the user has to later correct the orientation manually. The present system works in such cases; it first categorizes consumer images in a rotation invariant fashion and then detects their correct orientation. It is designed to be fast, using only low level color and edge features. A recently proposed information theoretic feature selection method is used to find most discriminant subset of features and also to reduce the dimension of feature space. Learning methods are used to categorize and detect the correct orientation of consumer images. Results are presented on a collection of about 7000 consumer images, collected by an independent testing team, from the internet and personal image collections.