Multi-Lingual Character Recognition and Extraction using Recurrent Neural Networks
A Novel Approach to Smart Autonomous Monitoring of Indoor Plant Health Based on Leaf Color
Edge Preserving Image Denoising with Deep Networks and Ensemble Convolutional Neural Network Architectures
Neural Network Approach to Detect Renal Calculi
Enhancing Fabric Quality Control: Implementing Real-Time Defect Detection with Image Processing Techniques and Arduino
Identification of Volcano Hotspots by using Resilient Back Propagation (RBP) Algorithm Via Satellite Images
Data Hiding in Encrypted Compressed Videos for Privacy Information Protection
Improved Video Watermarking using Discrete Cosine Transform
Contrast Enhancement based Brain Tumour MRI Image Segmentation and Detection with Low Power Consumption
Denoising of Images by Wavelets and Contourlets using Bi-Shrink Filter
Human emotions play an important role in decision making. Facial expression is natural and one of the most powerful immediate means for human beings to communicate their emotions and intentions. The face can express emotion sooner than people verbalize their feelings.As technology is advancing, the demand of face emotions recognition is increasing day by day for decision making. In past years, many researchers had introduced different techniques and different algorithms for accurate and reliable face emotion recognition. One key step in facial expression recognition is to extract the low dimensional discriminative features before the feature data are fed into classifier for emotions. In this paper, the authors have presented a new method of facial expression recognition based on statistical parameters of Qmatrix. The problem addressed here is to determine which features optimize classification rate so that, such features may be used in the evaluation of statistical features for face emotion recognition. The Q-matrix is much better than that of the remaining matrix methods, because it considers all possible neighbors of elements at once. The experimental results also demonstrate significant performance improvements due to the consideration of facial movement features and promising performance under face registration errors.
The brain is a highly dedicated organ. It serves as the control center for body functions. Words, events, thoughts, and feelings are centered in the brain. Each part of the brain has a specific, important function, often a great key function, and each part contributes to the healthy functioning of our body. The position of tumors in the brain shows its effect on individual's functioning and the symptoms caused by tumor. A color based segmentation technique that uses the kmeans clustering method is proposed to track the tumor objects in the Magnetic Resonance (MR) brain images. The key perception in color-based segmentation algorithm with K-means is to convert a given gray-level MR image into a color space image. Then the position of tumor objects is separated from other items of an MR image by using K-means clustering and histogram-clustering. In this paper, the K-means clustering algorithm is used for image segmentation and detecting the tumor objects that are found in the MR brain image.
In this paper, an unsupervised change detection method is proposed for satellite images and optical images. The multitemporal data and two discriminate areas of the land covers are changed between different time instances due to its robustness in opposition to noise. The UDWT is exploited to obtain a multi-resolution representation of the difference image which are taken from two satellite images acquired from the same geographical area but at different time instances. For multi-resolution representation of the difference image, a region-based active contour model is applied, then it is segmented into unchanged and changed regions. The proposed change detection method has been conducted based on the optical images. The extensive model results show that, the proposed change detection method constantly gives a better performance.
In this paper, a technique for enhancing the low resolution multi spectral image using Discrete Wavelet Transform (DWT) is proposed with the help of the corresponding high resolution panchromatic image. Image fusion, also called pansharpening, is a technique used to integrate the geometric detail of a high-resolution panchromatic (PAN) image and the color information of a low-resolution Multi spectral (MS) image to produce a high-resolution MS image. Remote Sensing systems, particularly those deployed on satellites, provide a redundant and consistent view of the Earth. In order to meet the requirements of different remote sensing applications, the systems offer a wide range of spectral, spatial radiometric and temporal resolutions. In general, sensors characterized with high spectral resolution, do not have an optimal spatial resolution, which may be inadequate to specific task identification in spite of its good spectral resolution. In panchromatic image with high spatial resolution, detailed geometric features can be easily recognized, while the multispectral images contain richer spectral information.
Image blur is a very difficult problem. Now a days, deblurring plays a vital role in digital image processing. Image deblurring has infinite solutions which are unstable because it is used to make the picture sharp and useful. So, to find the best solution for adjusting the regularization parameter, the number of iterations of the algorithm are used to address the optimization problem. This method well estimates the recovered image, but the residual image is a poorly deblurred image and it exhibits structured artifacts that are not spectrally white. Now in the proposed criterion, the same procedure mentioned above is repeated and the regularization parameter is chosen and the algorithm is decided to stop at the best Signal to Noise Ratio (SNR). Tests will be performed on monochrome and color images with various synthetic and real life degradations to get better results.
