Hybrid Speech-Based Text Processing System with Offline Recognition, Summarization, and Translation | IJCT Volume 13 – Issue 3 | IJCT-V13I3P106

This paper presented a Hybrid Speech-Based Text Processing System designed to perform speech recognition, multilingual translation, and text analysis in an integrated framework. The proposed system converts speech input into text, preprocesses the generated text, and translates it into multiple languages using neural machine translation models. Additionally, the system incorporates text processing techniques such as summarization and sentiment analysis to extract meaningful insights from the processed content. Unlike many traditional speech translation systems that rely heavily on cloud-based services, the proposed system supports offline speech recognition and processing, making it suitable for environments with limited internet connectivity. By integrating multiple modules into a unified pipeline, the system improves accessibility, supports multilingual communication, and enhances information understanding from spoken data. Experimental analysis demonstrates that combining speech recognition with neural machine translation and natural language processing techniques can provide an efficient and flexible framework for speech-based text processing applications. Future work may focus on improving translation accuracy for low-resource languages, optimizing model efficiency for faster offline processing, and extending the system to support additional languages and speech processing tasks.