Silvan Heller
Silvan Heller
This project lacks a license. The [resnet-converter](https://github.com/ry/tensorflow-resnet) has the MIT license but this one does not. This prevents people from uploading the converted model on their own site (with due...
This Draft PR contains work towards #370 and additionally has some cleanup for feature-related code. For testing purposes, there is a file named `external_extraction.json` which showcases the usage. This will...
Rerunning the `getExternalFiles` task can take a significant amount of time, it would be good to have a caching solution where if you run the task again, it recognizes already...
Currently, the `CottontailSelector` does not support batched NNS: https://github.com/vitrivr/cineast/blob/d132b193186e27a755cd0560361b7b8e0b3467da/cineast-core/src/main/java/org/vitrivr/cineast/core/db/cottontaildb/CottontailSelector.java#L253 There should at least be an unoptimized implementation, ideally this could be coordinated with the Cottontail devs.
Based on the work done in the [VBS 23 Branch](https://github.com/vitrivr/cineast/tree/VBS23), and to make vitrivr usable with retrieval features that are either not easily portable to java or only accessible externally,...
One user gave as a feedback to @cristinailli that they would like to be able to filter videos where no audio is present. This could be something we extract akin...
For some of the datasets commonly used with vitrivr, there might be items which should not appear in results when the general public can interact with demos. Instead of having...
Ideally, the Cottontail Access Layer would be refactored so that the schema to be used (currently `cineast`) can be specified in the config. This would mean we could write tests...
Currently, when using color sketches for a stage which is not the first one, it is possible that very little results are shown to the user. This is because for...
One thing we could extract akin to dominant color and duration, width or height (see e.g. the `TechnicalVideoMetadataExtractor`) is the aspect ratio of a video or image, which could be...