| 1.) |
SIH Hackathon 2023 |
SIH1409 |
Modern technology is changing rapidly and every individual has different level of understanding and different pace of learning things. Having same course for all personnel not only result in waste of time of departments and personnel's but also affects departments productivity. Proposed Solution: AI Based training system may be developed to real time design the course based on individual understanding and learning capacity. |
| 2.) |
SIH Hackathon 2023 |
SIH1418 |
Vision based methods using deep learning such as CNN to perform terrain recognition (sandy/rocky/grass/marshy) enhanced with implicit quantities information such as the roughness, slipperiness, an important aspect for high-level environment perception. |
| 3.) |
SIH Hackathon 2023 |
SIH1383 |
To develop a digital system that streamlines the appointment scheduling process in hospitals by automating the process of identifying doctor availability and appointment slot allocation. The system will utilize advanced technologies such as RFID, face detection, proximity of Mobile phone, or any other relevant technology to detect the presence of doctors in the hospital. The system will use Artificial Intelligence (AI) to allocate appointment slots based on the doctor's presence and the number of waitlisted patients. This will improve the overall patient experience by reducing the wait time. In conclusion, the proposed digital system will improve the efficiency and convenience of the appointment scheduling process in hospitals; the patients will benefit with reduced waiting time. |
| 4.) |
SIH Hackathon 2023 |
SIH1456 |
The problem entails transcription of audio files to the native script of the audio and then translation to English. The languages of interest are as mentioned below. A key element to be considered is that the solution would be required to be tuned for Indian accents. The languages of interest for the problem statement are shown below: ï‚· Hindi ï‚· Indian English ï‚· Urdu ï‚· Bengali ï‚· Punjabi Datasets pertaining to theses languages will be provided by us, which will consist of two major chunks of data: Training Set and a hidden test set. The participants will have only access to the Training set. They will develop their solutions based on the Training set. After the final solution submission by the participants, the final hackathon rankings will be decided by evaluation on the hidden test set. This is done to ensure that the participants solutions generalize better on newer data. The evaluation metric we want to use for this hackathon will be Word Error Rate (WER). The WER will be computed between the actual translated text with the solution generated text. The lower the WER the better the model. |
| 5.) |
SIH Hackathon 2023 |
SIH1460 |
All India Council for Technical Education (AICTE) has implemented a large-scale program to assess and improve the skills of engineering students in India in AICTE Approved Technical Institutions across the country. In respect of analysis, each of the online assessments (including those testing academic skills and those testing higher order thinking skills) is used for comparing student outcomes. More specifically, each of the Online assessments will be considered as unidimensional, reliable, and good at measuring a range of student ability. The vast majority of the items should also demonstrate psychometric characteristics of students. A Questions bank is already prepared and random selection of questions has been selected for the assessment. However, for the further development PARAKH teams needs a web-based adaptive multiple-choice question (MCQ) testing system that delivers MCQ assessments to users over the internet and adapts the questions based on the user's performance and abilities. A Pre-assessment or initial knowledge assessment can be designed to gauge the user's baseline knowledge; an initial assessment may be conducted. This assessment can consist of a set of MCQs covering different topics or difficulty levels. Based on that result a final assessment can be prepared. The system employs an adaptive algorithm that analyzes the user's responses, time spent on each question and level of difficulty and number of time questions has been asked in the past and also for performance data from the pre-assessment to determine their knowledge level and proficiency in specific areas. The algorithm uses this information to select subsequent questions that appropriately assessment to the students. Based on the adaptive algorithm, the system selects the next set of MCQs from a question bank. The questions chosen may vary in difficulty and content based on the student's performance. For example, if the student answers a question correctly, the platform may present a more challenging question next, while an incorrect answer may result in an easier question. AI based question generation should be integrated in the portal so that no random question is picked from the database and overall performance will be low. |
| 6.) |
SIH Hackathon 2023 |
SIH1417 |
Design and Development of AI-ML based intelligent de- smoking/de-hazing algorithm for reproducing the real time video of the area under fire specifically for indoor fire hazards to aid the rescue operation. |
| 7.) |
Problem Statement: Sign Language Recognition for Real-Time Communication |
-- |
The Real-Time Sign Language Recognition project aims to develop an advanced and user-friendly machine learning system that can instantly recognize sign language gestures and translate them into text or speech. The primary goal is to create a seamless communication bridge between individuals who use sign language as their primary means of expression and those who do not understand sign language. By harnessing the power of machine learning and computer vision, this project seeks to enhance inclusivity and foster effective communication for the hearing-impaired community. |
| 8.) |
Problem Statement: Plant Species Identification via Leaf Images for Botanical Research and Conservation |
-- |
The Plant Species Identification via Leaves project aims to develop a robust and accurate machine learning model that can automatically identify plant species from images of their leaves. The project holds significant importance for botanical research, plant conservation efforts, and environmental monitoring, as it offers a non-invasive and efficient method for plant species identification. |