Asim Banerjee

A Morphology Based Approach for Car License Plate Extraction

Locating the car license plate in an image or video frame of a car is an important step in car license plate recognition/identification applications. This problem poses many challenges like location of license plate from images taken in poor illumination and bad weather condition; plates that are partly obscured by dirt and images that have low contrast. This paper presents a new morphology based method for license plate extraction from car images. The algorithm uses morphological operations on the preprocessed, edge images of the vehicles. Characteristic features such as license plate width and height, character height and spacing are considered for defining structural elements for morphological operations. Connected component analysis is used to select the band containing license plate from the candidate segmented. The experimental results with a reasonably large set of car images are very encouraging.

Content Based Image Retrieval System for Multi Object Images Using Combined Features

This paper presents content-based image retrieval (CBIR) system for the multiobject images. The goal is to retrieve those images from the database, which contains the query object, which is a difficult problem when an image consists of multiple objects with arbitrary pose. In this paper we propose a method that uses GVF active contour for the efficient shape segmentation in a multiple object scenario. We have developed a novel technique for automatic initialisation of active contour. Using only shape information degrades performance in pose varying cases. As an attempt towards mitigating this problem, colour features are also used. The preliminary results obtained by our approach are very encouraging

Agro advisory system for cotton crop

In the agricultural domain, the main challenge is to present the new information and research to the farmers so that they can leverage the power of ICT to improve their agricultural practices and thereby the production. Huge amount of agriculture related data like weather data, soil health records, cropping pattern, location specific crop disease and pest are collected from different sources like services, remote satellites, and network of sensors. An agro advisory system presented in this paper helps to bridge the gap between farmers and the agriculture domain experts and developed for the cotton farmers in Gujarat region of India. The system consists of three basic components; Cotton Ontology, Web Services, and Mobile Application Development. The cotton ontology maintains domain knowledge required for answering farmer queries. The ontology contains information regarding crop, soil, cultivation process, disease, pest, and other relevant information. Protégé ontology development tool is used to develop this ontology. Appropriate Web services were built which help interactions with different data sources. The RESTful Web services are programmed in Java using the JAX-RS/Jersey API and the Eclipse EE IDE. The services are developed and deployed on a cloud based application server provided by Heroku. The Web services are invoked from the mobile device and in turn they connect to various data sources like Open Weather API, SQL database and the Ontologies. The farmers can use this application based on very simple android mobile interfaces. The prototype is developed using Java, Android SDK - v14 and Eclipse IDE.

A Theoretical Framework for Knowledge Management in Indian Agricultural Organizations

An effective agricultural knowledge management system can trigger continuous innovations in overall development of agriculture. It can also help in improving the livelihood of the rural communities in countries like India. The paper makes an attempt to develop a theoretical framework for understanding the flow of knowledge and its management in Indian agricultural organizations. This framework was developed based on inductive approach and a thorough review of existing frameworks of knowledge management (KM) process. The proposed framework is required for Indian agricultural organization because the existing frameworks are developed in the non-Indian and non-agricultural context. The proposed framework can be used by these organizations in better understanding and integrating both tacit and explicit knowledge.

A fast approach for edge preserving super-resolution

In this paper we propose a fast approach for edge preserving super-resolution (SR) based on learning of contourlet coefficients. Given a low resolution test image, we first obtain an initial HR estimate i.e., a close approximation to SR image by learning the contourlet coefficients from a training database consisting of low resolution (LR) and high resolution (HR) images. The final SR image is obtained by using a regularization framework in which both the SR and the LR images are modeled as separate homogeneous Markov Random Fields (MRFs). The LR image formation process is modeled as a decimated and noisy version of the SR image and the final cost function is minimized by using a gradient descent method. Novelty of our approach lies in preserving the edges in the final SR image while using a non edge preserving MRF prior. This is definitely advantageous since it avoids the use of discontinuity preserving prior and hence the computationally taxing optimization methods. The edges in the final SR correspond to those learned from the initial HR estimate. The use of MRF on the low resolution image imposes an additional constraint on the final solution and hence we expect a better solution. In addition, we use the initial HR image for estimating the decimation matrix entries as well as for learning the corresponding MRF parameter. We show the effectiveness of the proposed approach by conducting the experiments on images captured using a real camera.

Information Slicing: An Application to Object Classification in Satellite Images

Classification procedure aims at finding regions of the classes in the feature space. There are several algorithms proposed with supervised and unsupervised strategies for classification in literature. This paper goes on to propose a supervised method of classification using information slicing. Information lies in the feature space of the data to be classified. Training stage consists of slicing of the information by continuous partitioning of the feature space to find pure regions for classes from training data set. Classification then becomes a find-and-assign problem in the feature space for the input data. It has been shown here that this method of classification works well on data which have highly overlapping regions amongst the classes. This classification method is further applied for object classification in satellite imagery, where objects are homogeneous regions in images of considerably high resolution. Identification of homogeneous regions is done by a graph based segmentation algorithm which searches for segments having high intra region pixel similarity and high inter-region pixel dissimilarity. Feature vectors for each of these objects are calculated and then these feature vectors are given as an input to the information slicing classifier. Ground truth however is needed for the training stage of the classifier which has been generated here by manual intervention. Experiments have been done on four classes and results show that the method of information slicing works well on selective classes.

A fractal based approach for face recognition

The use of fractal technique known as iterated function systems (IFS) is mostly restricted to image compression till date. In this paper, we propose an automated face recognition system that operates in the fractal domain. This leads to an advantage towards handling large database of face images, which are compactly stored in terms of fractal codes. It is shown that candidate images of face recognition system could be recognized, efficiently, using interdependence of pixels arising from fractal codes (IFS) of images. The interdependence of the pixels is inherent within the fractal code in the form of chain of pixels. The preliminary results obtained by this method seem encouraging.

Alias minimization of 1-D signals using DCT based learning

In this paper, we propose a learning based approach for alias minimization of 1-D signals. Given an under-sampled test speech signal and a training set consisting of several speech signals each of which are under-sampled as well as sampled at greater than Nyquist rate, we estimate the non-aliased frequencies for the test signal using the training set. The learning of non-aliased frequencies corresponds to estimating them using a training set. The test signal and each of the under-sampled training set signal are first interpolated to the size of The non-aliased signals. They are then divided into a number of segments and discrete cosine transform (DCT) is computed for each segment. Assuming that the lower frequencies are non-aliased and minimally distorted, we replace the aliased DCT coefficients of the test signal with the best search from the training set. The non-aliased test signal is then re-constructed by taking the inverse DCT. The comparison with the standard interpolation technique in terms of both subjective and quantitative analysis indicates better performance.

Decimation Estimation and Linear Model-Based Super-Resolution Using Zoomed Observations

In this paper we present a model based approach for super-resolving an image from a sequence of zoomed observations. From a set of images taken at different camera zooms, we super-resolve the least zoomed image at the resolution of the most zoomed one. Novelty of our approach is that decimation matrix is estimated from the given observations themselves. We model the most zoomed image as an autoregressive (AR) model, learn the parameters and use in regularization to super-resolve the least zoomed image. The AR model is computationally less intensive as compare to Markov Random Field (MRF) model hence the approach can be employed in real-time applications. Experimental results on real images with integer zoom settings are shown. We also show how the learning of AR parameters in subblocks using Panchromatic (PAN) image gives better results for the multiresolution fusion process in remote sensing applications.

Decimation estimation and super-resolution using zoomed observations

We propose a technique for super-resolving an image from several observations taken at different camera zooms. From the set of these images, a super-resolved image of the entire scene (least zoomed) is obtained at the resolution of the most zoomed one. We model the super-resolution image as a Markov Random Field (MRF). The cost function is derived using a Maximum a posteriori (MAP) estimation method and is optimized by using gradient descent technique. The novelty of our approach is that the decimation (aliasing) matrix is obtained from the given observations themselves. Results are illustrated with real data captured using a zoom camera. Application of our technique to multiresolution fusion in remotely sensed images is shown.