http://dlib.iit.ac.lk/xmlui/handle/123456789/1613 2026-04-07T10:39:57Z http://dlib.iit.ac.lk/xmlui/handle/123456789/2294 Speechpal: Intergrating Automatic Speech Recognition and Large Language Models for Enhancing Expressive Communication in Aphasic Individuals. Gunathilaka, Pahan "This study introduces “SpeechPAL”, a speech recognition and prediction system designed for individuals with expressive language disorders. Throughout the project, the author has used Broca’s aphasia as an example of expressive language disorder, a disorder resulting from a neurological condition. The author utilizes Automatic Speech Recognition (ASR) and Large Language Models (LLMs) to develop the system and it serves as a tool for enhancing expressive verbal communication. In addition, during the testing and evaluation phases, the whisper model was assessed using the aphasia bank dataset and benchmarked. The chosen LLM could achieve a perplexity of 1.90, indicating the effectiveness of the model in predicting speech within the context of aphasia. " 2023-01-01T00:00:00Z http://dlib.iit.ac.lk/xmlui/handle/123456789/2216 Job Recommendation System Using Sentiment Analysis & Preference Adjustment Chandrasekaran, Rathushan "The research suggests a content-based job recommendation system to solve this issue by calculating the employer's sentiment score using sentiment analysis methods. The sentiment score is determined using reviews and ratings submitted by the employer company's current and former employees. The suggested system makes use of NLP techniques like data preprocessing, Vader Lexicon Score calculation, word-based sentiment analysis, and a Support Vector Machine Classifier model. The model was chosen from a pool of artificial intelligence algorithms trained on the English Google News 7B corpus. In order to rank job recommendations, the sentiment score is then combined with the similarity matrix and pairwise distance to calculate a weighted ensemble score. Based on user evaluations and the evaluation matrix for the developed algorithms, the prototype of the proposed system produces usable results. This research has successfully implemented a novel approach that integrates sentiment analysis to check employer credibility in job recommendation systems. The proposed system offers useful information that helps job seekers choose their careers wisely." 2023-01-01T00:00:00Z http://dlib.iit.ac.lk/xmlui/handle/123456789/2215 Detection and Quantification of Potato Late Blight Using Object Detection and Instance Segmentation Vidanagamage, Janith "The detection of Potato late blight (PLB) disease is a challenging task for farmers, who currently rely on traditional methods such as visual observation that are usually slow and subject to human error. This has led to the quick spread of the disease and significant yield loss. Adding on to that, due to wrong judgments made by farmers, fungicide wastage has also been increased which has led to an increase in expenses over the years. Further investigations found that plant pathologists in the country face another serious problem when quantifying the PLB disease. Even though a protocol exists to quantify the disease, due to the variations in individual experience in the field, the quantification of the disease will differ from one pathologist to another. A faster, more accurate solution is needed to help both farmers and pathologists to detect and quantify PLB disease early so that it will be possible to control its spread and reduce yield loss. A novel approach has been suggested in this study to build a solution for the above problem. This approach employs object detection and instance segmentation techniques to identify and quantify the disease. This will be solved in two stages. Stage one will use a deep learning model based on the Mask R-CNN architecture to extract leaf instances from an input image which also includes complex backgrounds such as soil, weed plants. For the second stage, another similar model based on the same architecture will be built to segment the healthy and diseased regions from the extracted leaf instances. Afterwards a calculation will be made based on the total heathy and unhealthy area of the identified leaves to quantify the disease. In the first stage, the novel model was trained on a dataset using different types of augmentation techniques, achieving a mean average precision (mAP) of 75%. In the second stage, a model was trained using masks obtained from the stage 1 model, yielding a mAP of 81%. The severity calculation was made based on the pixel area of both healthy and unhealthy areas of the leaf instance. The study found a higher correlation between the manual severity values and the value generated by the proposed model. The results demonstrate the potential of the proposed model in accurately detecting and assessing the severity of plant diseases, with implications for precision agriculture and sustainable farming practices." 2023-01-01T00:00:00Z http://dlib.iit.ac.lk/xmlui/handle/123456789/2214 Tamil Aspect Base Sentiment Analysis for Movie Recommendation System Sriranjan, Abilash "This research will be based on extracting the characteristics from the tamil reviews/texts provided for a product/service by the existing users to be used to find sentiment score for the feature or attribute extracted and recommending the user by providing a straightforward idea on what aspects of the service/product are preferred by the user. There are various Recommendation systems available in the market. Some of them have used tamil dataset for their recommendation system. Through this research recommendation accuracy for recommendation systems which use tamil data sets can be increased. Extracting the feature from the tamil review and giving polarities for the whole product / service as well as for the extracted feature would give the user a clear vision about the product/ service and accuracy of recommendation for the product or a service will also be increased." 2023-01-01T00:00:00Z