Search

In today's era, video analysis is immensely involved in recognizing the sport-related movement that has become a significant part of human's life. The intent of this approach is to know about the player's activities with prior information of tracking objects. It also analyzes the player potential or capacity to lead the winning team. When the player frequently changes their location, object tracking and action recognition will become a quite challenging task. Over the game, various athletes or different objects are considered to assist the system to easily recognize the respective actions of the player. Most of the previous models have been implemented, yet, it faces such consequences to provide promising performance. To meet the pre-requisite, a new multi-athlete tracking model for action recognition in soccer sports is designed with deep learning approaches. Initially, the multi-object tracking video is offered as the input to pre-processing phase. Here, occlusion and background clutter removal and contrast enhancement techniques are utilized to perform pre-processing in the videos. Then, the pre-processed video is offered to the multi-object tracking phase, where the jersey number is observed during multi-object tracking to avoid the identity switch problem. Then, effective multi-object tracking is performed by adaptive YOLOv5. The parameters presented in the improved adaptive YOLOv5 are tuned by proposing a new algorithm as the Random-based Cheetah Red Deer Algorithm (RCRDA). Next, in the action recognition phase, the tracked object from the video is taken based on the Region of Interest (ROI) that is subjected to an action recognition model named Optimized Transformer-based Residual Network with Deep Convolutional Network (OTRN-DCN). At first, ROI is offered as the input to TRN for attaining the feature vectors. Then, the optimal weighted vector extraction is performed, where the weight is tuned by the developed RCRDA. Finally, the attained optimally weighted vectors are given to the DCN phase for attaining recognized action as output. Hence, the developed multi-object tracking and action recognition model will secure an improved recognition rate than the traditional framework. [ABSTRACT FROM AUTHOR]

[Display omitted] • There are a lot of abnormalities in the sizes and location of the brain tumor(s). This makes it really difficult for complete understanding of the nature of the tumor. • Often times in developing countries the lack of skillful doctors and lack of knowledge about tumors makes it really challenging and time-consuming to generate reports from MRI. • A novel feature extraction method to extract the basic features using deep learning models and radiomic features of 93 using pyradiomics to understand the nature of the tumor and classify it accordingly. • In the literature limited number of MRI images was used, in our work two benchmark datasets are utilized to improve the accuracy of the recommended model. • Novel filtering techniques, improved preprocessing methods and an enhanced hybrid optimized deep learning machine was developed are used to improve the efficiency of the suggested model. • Preprocessing using literature algorithm is complex and time consuming for noise removal. • The state-of-art depends on the Convolution operation for feature gathering, in our system a novel feature fusion was initiated to analyze the nature of the tumor for accurate classification. • A novel deep learning classifier was developed to classify the brain tumors into four specified categories. • The classification performance was measured using eight performance metrics. • To validate the performance of our proposed system, it was compared with the existing deep learning approaches stated in the paper with respect to the accuracy parameter. • The overall performance and comparison of our system outperforms the literature Brain Tumor Classification and Brain Tumor (MRI) datasets. Brain Tumor classification in MRI images is a time-consuming and tedious task for medical professionals. An accurate classification model can assist healthcare providers in treating patients with effective care. In this research work, an enhanced learning machine for classifying brain tumors has presented for medical specialist's assistance. Deep learning architectures like Inception V3 and DenseNet201 are used to retrieve the categorization model's basic features. Along with the features collected using deep learning models, radiomic properties are integrated before classification in order to increase classification accuracy. Particle swarm optimized kernel Extreme Learning Machine (PSO-KELM) model has used to categorize the features into four groups like No Tumor, Gliomas, Meningiomas and Pituitary Tumors. Our system employs two benchmark datasets to evaluate the efficiency of the developed classification system using measures such as accuracy, recall, precision, false-positive rate, recall, precision, f1-score, and AUC ROC score, all of which our model performs better than the literature values. In addition, the four existing optimized learning methods are individually compared with dataset 1 and five approaches are independently evaluated with dataset 2. Accuracy measure is used to authenticate the improved performance analysis of oursuggestedsystem. In training and testing phases, the suggested model accomplishesimproved accuracy than theState-of-Art deep learning approaches. Our system's classification accuracy is 96.17% and 97.92% for datasets 1 and 2, respectively. Similar to the training method, the proposed testing model's accuracy is improved as 97.97% and 98.21%, respectively. [ABSTRACT FROM AUTHOR]