Adding Tools - Captioning, Image Processing, Generic OD & Seg

[shankar-vision-eng]

This PR includes the following features

• Adds capability to extract_frames to extract frames from youtube videos
• Adds Florancev2 image captioning in two modes - Brief and Detailed.
• Adds Generic OD and Segmentation tool to detect common objects with text prompts
• Adds Florancev2 based image question answering with context
• Adds following image generation/processing tools
• • Image to Depth - Use Depth Anything V2
• • Image to Pose - Generates Open pose stick image which can interpreted by Open pose libs
• • Image to Soft Edge Image - Generates HED or Soft edge image of object boundaries
• • Image to Normal - Generates Normal image where RGB values represent XYZ surface normals
• Adds template matching tool to find cropped ROI in original image

[shankar-vision-eng]

PR looks good, but can you use the model names in to tool function names? For example image_question_answering_with_context and the generic tools

Renamed tools with actual model names