Chronic heart disease (CHD) is a widespread and persistent health challenge that demands immediate attention. Early detection and accurate diagnosis are essential for effective treatment and management of this condition. To overcome this difficulty, we created a state-of-the-art IoT-Based Ambulatory Blood Pressure Monitoring System that provides real-time blood pressure readings, systolic, diastolic, and pulse rates at predefined intervals. This unique technology comes with a module that forecasts CHD’s early warning score. Various machine learning algorithms employed comprise Naïve Bayes, K-Nearest Neighbors (K-NN), random forest, decision tree, and Support Vector Machine (SVM). Using Naïve Bayes, the proposed model has achieved an impressive 99.44% accuracy in predicting blood pressure, a vital aspect of real-time intensive care for CHD. This IoT-based ambulatory blood pressure monitoring (IABPM) system will provide some advancement in the field of healthcare. The system overcomes the limitations of earlier BP monitoring devices, significantly reduces healthcare costs, and efficiently detects irregularities in chronic heart diseases. By implementing this system, we can take a significant step forward in improving patient outcomes and reducing the global burden of CHD. The system’s advanced features provide an accurate and reliable diagnosis that is essential for treating and managing CHD. Overall, this IoT-based ambulatory blood pressure monitoring system is an important tool for the early identification and treatment of CHD in the field of healthcare.

Region of interest means area of image that is to be selected or considered as important part of an image. Region of interest is selected in two different ways; first region is selected by user and second is system dependent selection. Instead of using system selection, user selection will be more effective because system selection can create difficulties while considering regions whereas user will select particular region in user selection. Query images when provided by user its global features are extracted but the problem will be with the local features, which are not considered during image retrieval. In order to solve the problem of local features extraction, region of interest within particular image is introduced and instead of retrieving whole image desired region of the images are focused for image retrieval. This paper gives the idea of region of interest based image retrieval concept and is expected to be helpful to the researchers who are working in region based image retrieval systems.

Logo is used for identity purpose and logo matching is one of the challenge in the recent era in a various applications to meet customers expectations for a product or to discover improper use of logo. Most of the author contributed in this domain, traditionally edge detector and shape detectors are used for this purpose. But still there are some difficulties in logo matching like different images contains logo with different shapes and size, location of logo is not fixed in each image and occlusion problem etc. For these types of logo, traditional methods become fail. To overcome the problem associated with traditional method in this paper literature survey of previous and recent work is presented, in which different descriptors i.e SURF, SIFT, HOG, methods and hybrid version of descriptor i.e ORB used for logo matching. This paper presents a concise review on logo matching and recognition and expected to be useful for researchers who working both on logo matching and object recognition, providing a most comprehensive and accessible information of this important field of research.

Clothing is important to daily life. As on-line fashions retail industry has been growing recognition and retrieval of relevant clothing items have huge potentials, because of the highly increasing earnings of the online clothing sale. This involves acquiring relatable clothing data from a huge number of databases available online. Extracting knowledge and important insights from fashion data still presents challenges due to the essential subjectivity needed and lack of influence to effectively model the domain. There have been many researches about clothing retrieval which are assisted by automated system. In fashion industry, various techniques are suggested and implemented for the classification, recognition and extraction of relevant images from huge databases. The success rates of these techniques are judged on various parameters like their accuracy rate, time taken to generate the outputs, etc. The paper presents a brief analysis of the various retrieval techniques used for clothing retrieval in the fashion industry. Along with this a complete summary on the existing techniques of retrieval based on different scenarios is explained which is expected to be helpful to the researchers who are working in fashion domains for image retrieval.