IJIRCSTJournal International Journal of Innovative Research in Computer Science and Technology I S Volume 10 Issue 6 Electronics and Communication Engineering English November - December 2022 N N Y 2022 11 07 Computer Sciences A New Residual Convolutional Neural Network-Based Speech Improvement English Y 39 42 M. Balasubrahmanyam English Y B. Haribabu English N P. Uday Kumar English N https://doi.org/10.55524/ijircst.2022.10.6.7 Among the most crucial methods for denoising a noisy voice signal and enhancing its quality is speech enhancement. This study makes use of Adaptive Residual Neural Network technique to reduces maximum off background noise. This method continuously monitors the background noise depends upon the environmental changes using SNR parameter. It has two functions first one is non linear functions followed by convolutional neural networks and second one is linearity followed due to Residual neural networks. By using these factors remove background noise even SNR is low conditions. Compared to other techniques this technique is new,fastest, requires less training and also reduces size.   English Convolutional Neural Networks, Liver, Segmentation, Tumor, ResNet, Deep Learning. https://ijircst.org/abstract.php?article_id=1051