Daily Archives: May 8, 2025

Stock Predator: ML-driven Stock Prediction
Authors:-Anushka Sakure, Shrishti Mishra, Riya Das, Reetika Roy

Abstract-:Stock price prediction remains a challenging task due to the inherent volatility and non-linear nature of financial markets. This study proposes a deep learning approach using Long Short-Term Memory (LSTM) networks to forecast stock prices, leveraging their ability to model temporal dependencies. Historical data from the S&P 500 index (2010–2023) was pre-processed, normalized, and used to train an LSTM model. The model’s performance was evaluated against ARIMA and SVM using RMSE, MAE, and directional accuracy. Results indicate that the LSTM model outperforms traditional methods, achieving an RMSE of 1.82 and 87% directional accuracy. This work highlights the potential of LSTM in financial forecasting and algorithmic trading strategies.

DOI: 10.61137/ijsret.vol.11.issue2.433

Published by: Kajal Tripathi

Adventure of Artemis (2D Game for PC using Unity Engine)
Authors:-Vaibhav Singh, Lucky Yadav, Shivam Dewangan, Shubham Singh, Professor Ravikant Soni

Abstract-:This project documents the collaborative work of four individuals in the creation of a 2D game, using the Unity Engine, C# and Visual Studios. Merging tech skills with natural creative talent, the team is on a mission to build an unforgettable and enjoyable gaming adventure. The project unfolds as a testament to the quality of game development, exploring the combined contributions of programming, artistry, and design. By leveraging modern game development tools and technologies, the team moves through all the complexities of game mechanics, level design, background score integration and wonderful sound effects. The culmination of their efforts is a polished arcade styled game that exemplifies their collective dedication, innovation, and expertise. Through this project, the team presents a comprehensive narrative of their collaborative journey, offering insights into the challenges, triumphs, and lessons learned in the pursuit of gaming excellence.

DOI: 10.61137/ijsret.vol.11.issue2.432

Published by: Kajal Tripathi

Crafting Worlds: 3d Animation
Authors:-Agarwal Sneha, Anampaka Sneha, Repalle Deepika Persis, Assistant Professor Mrs. Nandita Manvar

Abstract-:The process of rebuilding 3D high-resolution (HR) [1]models from 2D photographs has grown very important in multiple applications, such as augmented reality (AR), virtual reality (VR) [22], games[6] , medical imaging [12], and digital content creation. Conventional methods of 3D scanning may need costly equipment, making access difficult. The current study demonstrates an AI-based system that fully automates 3D model generation from 2D photographs based on computer vision and deep learning methods. The system utilizes Neural Radiance Fields (NeRF) [11], Open3D[8], OpenCV [18], and Blender API to generate high-quality 3D reconstructions. Images are uploaded through a web interface by the users, which are then processed through a pipeline of sparse structure generation, structured latent generation, and multi-image optimization. Exporting the resulting models in different formats like Gaussian Splats and GLB is supported by the system, which allows them to be used in different applications. Furthermore, a 10-second animated visualization is created based on OpenCV[18] and FFmpeg[16] to increase user interaction with the model. The suggested method presents an improved multi-image processing algorithm that enhances depth reconstruction and estimation accuracy. In contrast to conventional photogrammetry techniques that are plagued by perspective changes and uneven lighting, this technique corrects 3D models through structured latent learning. Future developments are real-time rendering via WebGL or Three.js [14], cloud processing for scalability, greater file format compatibility, and AI-based texture augmentation. Further functionalities like AR/VR integration, automatic animation synthesis, and a marketplace with community support will make the platform even more usable. By offering a scalable and user-friendly solution, this work closes the gap between sophisticated 3D modelling technologies and practical applications, enabling wider use in fields that need precise 3D reconstructions.

DOI: 10.61137/ijsret.vol.11.issue2.431

Published by: Kajal Tripathi