Shri Andal Alagar College of Engineering

Skin cancer is usually classified as melanoma and non-melanoma. Melanoma now represents 75% of humans passing away worldwide and is one of the most brutal types of cancer. Previously, studies were not mainly focused on feature extraction of Melanoma, which caused the classification accuracy. However, in this work, Histograms of orientation gradients and local binary patterns feature extraction procedures are used to extract the important features such as asymmetry, symmetry, boundary irregularity, color, diameter, etc., and are removed from both melanoma and non-melanoma images. This proposed Efficient Classification Systems for the Diagnosis of Melanoma (ECSDM) framework consists of different schemes such as preprocessing, segmentation, feature extraction, and classification. We used Machine Learning (ML) and Deep Learning (DL) classifiers in the classification framework. The ML classifier is Naïve Bayes (NB) and Support Vector Machines (SVM). And also, DL classification framework of the Convolution Neural Network (CNN) is used to classify the melanoma and benign images. The results show that the Neural Network (NNET) classifier’ achieves 97.17% of accuracy when contrasting with ML classifiers.

The high death rates are occurred due to the Melanoma among the skin tumor persons. Melanoma is more dangerous when it raises inside of the skin layer. Hence, watch the wound in depth of the skin is a significant cause to identify melanoma. A (NI) non-invasive computerized dermoscopic (DS) method is introduced in these study. Existing DS system many faces various challenges includes segmentation and classification for detecting the skin cancer. The objective of the research work to improve the segmentation and classification performance. In DS images hair removal and segmentation are performed by using Hybrid Laplacian of Gaussian (HLOG) filter and Flexible Kernel-Based Fuzzy Means (FKFCM), whereas Patch-Local binary patterns (LBP) for feature extraction. The extensive experiment are conducted on largest publicly available benchmark dataset such as PH2, Kaggel and HAM 10000. To validate the performance of proposed technique when compared with traditional segmentation and classification techniques. The proposed system archive 97% of accuracy, 98% sensitivity and 96% of specificity for PH2 dataset. The planned scheme stands out among the few modern literary sources presented in the context of the analysis of DS images in terms of productivity and accepted methodologies, which proves the reliability of the novel study.

In recent decadesmany research works have been investigate the benefits of the mobile sink node enabled energy aware WSN routing to improve the throughput rate of the sensor networks. It is well proved that throughput rate drivensensor networks models are highly sensitive to delay-bound applications, where all information's are gathered are need to be forwarded within a given bounded delay. In generalenergy conservations are largely depends on the tour path taken by the mobile sink. Here in order to optimize the mobile sink tour path a hybrid approach is proposed to regulatethe travelling path and also accommodate the dynamic location identification where the mobile-sink node will locate only in specified sensor node which is called primary node (PSN), as opposed to travelling all possible nodes in WSN. Information gathered from all other nodes are collected in PSN nodes through local routing and forward their packets via one hop distance to the nearest PSN. Then two basic fundamental problems are arises: formulation of mobile sink travelling paths that can visit all PSN's within a given time delay and optimizing the energy consumption of mobile sink to improve the overall life time of the WSN. To address the primary concern problem, a heuristic called adoptive primary sensor node (PSN) selection is proposed, where each sensor node is connected to corresponding to PSN node to its hop distance from the mobile sink path taken and the traffic rate that forwards to the closest PSN. And energy aware local routing is used for energy efficiency. Finally PSN based mobile sink model is validated via extensive network simulation, and through simulation results we proved that proposed energy aware PSNmodel allows mobile sink to gather all local information within a given time while. Local routing for each PSN node reduces energy consumption by 39% and increases network lifetime by 12 %, as compared with all other state-of-the-art algorithms.

Aim: Aquatic system poses a serious threat from industrial effluents having heavy metals and chromium due to its wide use in industry. It becomes a prime responsibility for removal of chromium ions from effluents before discharged into the aquatic body. Biosorption comes as an effective technology in the removal of heavy metals from industrial effluents. This study deals with biosorption of chromium from aqueous solution by Chaetomorpha antennina. Methodology: Biosorption of chromium was studied using microalgae Chaetomorpha antennina, as an adsorbent, which was characterized using Fourier Transform Infrared (FTIR) and Scanning Electron Microscope (SEM) to determine the functional groups and the structural characteristics, respectively. Equilibrium and optimization studies for various parameters like pH (1 - 11), adsorbent dosage (0.2 - 1 g l(-1)), initial chromium concentration (10 mu g - 100 mu g ml(-1)) and agitation time (0 - 50 mins) were examined using batch process. Results: The maximum percent removal of chromium ions was found to be 83% at pH 1 at an adsorbent dosage of 1g at a contact time of 30 min, which showed maximum adsorption. Interpretation: Biosorption of chromium ions onto the surface of Chaetomorpha antennina, showed that biosorption was dependent on the equilibrium pH of the solution, biosorbent and contact time.

Compression of medical images is essential for teleradiology applications. Medical image compression saves time and bandwidth during data transmission and also reduces storage space. For medical images, the diagnostically useful information termed Region of Interest (ROI) is localized in a small area. A compression algorithm to preserve high quality in diagnostically significant regions and allowing degradation in other regions providing higher compression is necessary. In this proposed work, ROI is processed using lossless compression algorithm and lossy compression elsewhere in the image. Variational level sets with distance regularization segment ROI. The compression algorithm is developed to obtain image with high optimum compression efficiency and also with high fidelity, especially for ROI. A new threshold based medical image compression with edge preservation is proposed to preserve diagnostic information within ROI. The performance of the proposed ROI based compression algorithm is compared in terms of Compression Ratio (CR) and Peak Signal to Noise Ratio (PSNR).