P. V. V. Kishore

Visual-verbal machine interpreter for sign language recognition under versatile video backgrounds

This research paper is an attempt to create a video background independent sign language recognition (SLR) system. SLR acts as a Machine Interpreter (MI) between a mute person and normal person. One of the key difficulties in sign language recognition is video background of a sign video in which signer is located. Signer is extracted from cluttered back video backgrounds using boundary and prior shape information. Active contours energy function is formulated by amalgamating energy functions from boundary and shape prior elements. Energy minimization for movement of active contour is achieved using Euler- Lagrange equations. Feature vector is constructed from the segmented signer frames using a frame average based pooling function along with the shape inform obtained from active contour. Artificial Neural Network is constructed to classify and recognize gestures from video frames of signers. Compared to traditional methods of sign language recognition, the proposed Visual-Verbal Machine Interpreter (V2MI) for sign language recognition offers a recognition rate of around 93%.

4-Camera model for sign language recognition using elliptical fourier descriptors and ANN

Sign language recognition (SLR) is considered a multidisciplinary research area engulfing image processing, pattern recognition and artificial intelligence. The major hurdle for a SLR is the occlusions of one hand on another. This results in poor segmentations and hence the feature vector generated result in erroneous classifications of signs resulting in deprived recognition rate. To overcome this difficulty we propose in this paper a 4 camera model for recognizing gestures of Indian sign language. Segmentation for hand extraction, shape feature extraction with elliptical Fourier descriptors and pattern classification using artificial neural networks with backpropagation training algorithm. The classification rate is computed and which provides experimental evidence that 4 camera model outperforms single camera model.

Image enhancement using virtual contrast image fusion on Fe3O4 and ZnO nanodispersed decyloxy benzoic acid

Low contrast and noisy photographic pictures can be considerably improved by image-processing techniques. Techniques like histogram equalisation produce high-contrast images but often fail to preserve the colour texture information. To overcome this deficiency, a contrast-enhancement approach has been devised, using virtual contrast image fusion (in Haar wavelet domain). This technique has been evaluated by a study of the optical textures of nano-dispersed decyloxybenzoic acid with small quantities of Fe3O4 and ZnO added.

Medical image watermarking using RSA encryption in wavelet domain

This research highlights the use of encryption based medical image watermarking of three types of medical images namely, MRI, CT and Ultrasound. The watermark used in this application is patient image. The watermark patient image is encrypted using a public Key encryption with RSA algorithm. Cover image is a medical image of standard resolution to be watermarked. The RSA encrypted patient image is embedded into the medical cover image in wavelet transform domain. Inverse wavelet transform is applied to produce a watermarked medical image with patient image as watermark. The watermarked image is tested in both visual quality and quantitatively using psnr and normalized cross correlation(ncc). The extraction processing is done with the help of Private Key in wavelet domain. The algorithm was tested for all three types of medical images and using db2 mother wavelet for level-2 decomposition. Attacks are planned on the extraction process and encrypted watermark is extracted successfully.

Train rolling stock segmentation with morphological differential gradient active contours

Rolling examination as it is called by railway maintenance staff of Indian railways is visual and auditory examination of moving bogies of a train for defects. The undercarriage moving parts of the train are called rolling stock. This paper makes an attempt to segment the rolling stock from video frames of the rolling stock for further analysis. This paper focuses on Chan vese active contour (CV) model for segmenting the rolling stock. We present a modified version of Chan vese using morphological differential gradient (CVMDG) to segment rolling stock. The rolling stock videos are captured under four different lighting conditions near Guntur railway station in INDIA. For better segmentation of rolling stock, video frames are contrast enhanced with virtual exposure wavelet image fusion. The segmented rolling stock is compared with ground truth model to assess the usability of the proposed method for rolling stock segmentation.

Block based thresholding in wavelet domain for denoising ultrasound medical images

Medical ultrasound imaging has transformed the disease identification in the human body in the last few decades. The major setback for ultrasound medical images is speckle noise. Speckle noise is created in ultrasound images due to numerous reflections of ultrasound signals from hard tissues of human body. Speckle noise corrupts the medical ultrasound images dropping the detectable quality of the image. An endeavor is made to recover the image quality of ultrasound medical images by using block based hard and soft thresholding of wavelet coefficients. Medical ultrasound image is transformed to wavelet domain using debauchees mother wavelet. Divide the approximate and detailed coefficients into uniform blocks of size 8×8, 16×16, 32×32 and 64×64. Hard and soft thresholding on these blocks of approximate and detailed coefficients are applied. Inverse transformation to original spatial domain produces a noise reduced ultrasound image. Experiments were conducted on medical ultrasound images obtained from diagnostic centers in Vijayawada, India. Quality of improved images in measured using peak signal to noise ratio (PSNR), image quality index (IQI), structural similarity index (SSEVI).

Medical image watermarking with DWT-BAT algorithm

Medical images communicate imperative information to the doctors about a patients health situation. Internet broadcasts these medical images to inaccessible sites of the globe which are inspected by specialist doctors. But data transmissions through unsecured web invoke validation problems for any image data. Medical images that are transmitted through the internet must be watermarked with patient pictures for substantiation by the doctors to ascertain the medical image. Watermarking medical images necessitate attentive adjustments to protect the information in the medical images with patient image watermarks. The medical images are used as an envelope image in the watermarking process which is visible on the network. These envelope medical images are watermarked with patient images in wavelet domain there by using the BAT algorithm form optimizing the embedding process for peak signal to noise ratio (psnr) and normalized cross correlation coefficient (ncc) values. The medical image envelope and letter inside envelope i.e. watermark image are transformed into wavelet domain and are mixed using scaling factor alpha which is termed as embedding strength. BAT algorithm is an optimization algorithm specialized in optimizing the values of peak-signal-to-noise ratio for a particular value of alpha, the embedding watermark strength. Finally these watermarked medical images are put on the network along with the secret key that will be used for extraction. At the receiving the embedded watermark is extracted using 2DWT using the embedding strength value using BAT algorithm. The robustness of the proposed watermarking techniques is tested with various attacks on the watermarked medical images. Peak-Signal-to-Noise ratios and Normalized cross correlation coefficients are computed to accesses the quality of the watermarked medical images and extracted patient images. The results are produced for three types of medical images with one patient image watermarks using single key by using four wavelets (haar, db, symlets, bior) at four different levels.

Optical Flow Hand Tracking and Active Contour Hand Shape Features for Continuous Sign Language Recognition with Artificial Neural Networks

To extract hand tracks and hand shape features from continuous sign language videos for gesture classification using backpropagation neural network. Horn Schunck optical flow (HSOF) extracts tracking features and Active Contours (AC) extract shape features. A feature matrix characterizes the signs in continuous sign videos. A neural network object with backpropagation training algorithm classifies the signs into various words sequences in digital format. Digital word sequences are translated into text with matching and the suiting text is voice translated using windows application programmable interface (Win-API). Ten signers, each doing sentences having 30 words long tests the performance of the algorithm by computing word matching score (WMS). The WMS is varying between 88 and 91 percent when executed on different cross platforms on various processors such as Windows8 with Inteli3, Windows8.1 with inteli3 and windows10 with inteli3 running MATLAB13(a).

Image enhancement of nano-dispersed N-(p-n-decyloxybenzylidene)-p-n-hexyloxy aniline using combined unsharp masking

ABSTRACT The main objective of the image enhancement is to process an image with suitable technique to produce better visibility for a specific application. To identify key features like transition temperatures, clear phase identification in the liquid crystalline images, we require some novel image processing techniques. Characterisation and mesomorphic behaviour in pure and 1% ZnO nano-dispersed liquid crystalline N-(p-n-decyloxybenzylidene)-p-n-hexyloxy anilines, 10O.O6 compounds are carried out using a polarising microscope and images are preserved for enhancement. Both the compounds exhibits NACIG (nematic, smectic-A, smectic-C, smectic-I, smectic-G) phases and the transition temperatures of the 1% ZnO nano-dispersed 10O.O6 are reduced compared with pure 10O.O6. Further, in this paper, a novel image enhancement technique of combined unsharp masking is proposed on pure and 1% ZnO nano-dispersed 10O.O6 liquid crystalline compounds for better visibility of phases at transition temperatures. The proposed method is used to identify the uniform regions and to detect the defects which may not be clearly observed from the textures that are recorded by polarising microscope.

Crowd Density Analysis and tracking

Crowd Density Analysis (CDA) aims to compute concentration of crowd in surveillance videos. This paper core is to estimate the crowd concentrations using crowd feature tracking with optical flow. Local features are extracted using Features for Accelerated Segment Test (FAST) algorithm per frame. Optical flow tracks the features between frames of the surveillance video. This process identifies the crowd features in consecutive frames. Kernel density estimator computes the crowed density in each successive frame. Finally individual people are tracked using estimated flows. The drawback of this method is similar to suffered by most of the estimation methods in this class that is reliability. Hence testing with three popular optical flow models is initiated to find the best optical flow. Three methods are Horn-Schunck (HSOF), Lukas-Kanade (LKOF) and Correlation optical flow (COF). Five features extraction methods were tested along with the three optical flow methods. FAST features with horn-schunck estimates crowed density better than the remaining methods. People tracking application with this algorithm gives good tracks compared to other methods.