Tenders Are Invited For Computer Upgrade Services - Creation And Updating Of Applications, Systems, Websites ... in Slovenia

Work Detail

Computer Upgrade Services - Creation And Updating Of Applications, Systems, Websites And Platforms - Development Of The Etnoteka Application. The Following Activities Are Planned Within The Project*: A. Back-End Development: - Crud Database Development; - Modules For User Administration And Data Management; - Support For Llm And Lmm Vector Files, Which Will Enable Automatic Annotation And Metadata Generation Based On Video And Photo Analysis; - Database Export In A Standardized Format (Json, Xml, Xlsx/Csv), According To The Criteria Specified By The Administrator. B. Development Of The Front-End: B.1 Interfaces For Annotation Of Audiovisual Material: - Interface For Temporal Annotation Of Visual Content (Video Clips); - Support For Automatic Annotation Of Visual Content Based On Llm Or Lmm (Multilingual Models And Models Optimized For Slovenian And Dialects); - Possibility Of Manual Annotation Of Visual Content. B.2 Interfaces For Search And Playback: - Soft (Fuzzy Or Fusion Rank) Search, Via Several Categories And With Logical Operators; - Search By Time; - Search By Geolocation; - Thematic Search; - Semantic Search Of Video And Photos: Textual Descriptions Of Generative Ai (Based On Lmm And Llm Vector Embeddings); - Hybrid Video And Photo Search (Combination Of Fuzzy And Semantic Search); - Interface For Playing Visual Content. B.3 Interfaces And Mechanism Of Interaction With Lmm: - Interface For Communication With Generative Ai, Which Will Generate Annotations For The User (Prompt Engineering, Example Of Communication With Chatgpt). C. Data Processing, Machine Learning And Generative Ai C.1 Data Processing: - Processing Of Place Names To Generate Geolocations (Long, Lat); - Video Processing For The Preparation Of Vector Files (Segmentation, Formatting); - Text Processing For The Preparation Of Vector Files (Lemmatization, Tokenization). C. 2 Machine Learning And Generative Ai: Visual Material - Video Segmentation Based On Machine Learning: Development Of Video Segmentation Algorithms, For E.G. Capture Of Relevant Transitions And Activities Within The Video, Also Based On Mpeg Visual Features (E.G., Color Structure Descriptor (Csd), Edge Histogram Descriptor (Ehd), Fast Scene Change Detection Algorithm); - Implementation Of Large Multimodal Models (Lmm) With The Purpose Of Automatic Annotation And Semantic Search Of Vector Files Of Visual Content; - Optimization Of Models (Fine-Tunning) On ​​Visual Content And Contexts And With The Help Of Taxonomies Of Folk Phenomena; - Calculation Of Lmm Vector Submissions And Entry Into The Database; Text - Implementation Of Large Language Models (Llm) For Slovenian With The Purpose Of Semantic Search By Vector Embeddings; - Optimization Of The Model (Fine-Tunning) On ​​Concrete Texts; - Calculation Of Lmm Vector Embeddings And Entry Into The Database. D. Design: - Ux Flow (Plan Of The Applications Appearance, Functionality And Interactions); - Designing The Application, Search Engines And Individual Elements For Displaying Content; - Responsive Design (Adapting To Different Screens/Devices); - Support For Major Browsers (Firefox, Safari, Chome). E. Placement On A Virtual Website

Contact Information