1.4.0 - Extra usage of Google Maps when other options fail

• When neither Google Vision, Azure Vision, Google Maps (with word search) or Bing Maps (with word search) return any results, use the full Google Maps results (without any word search) and define landmark based on GPS proximity to photo.
• Improvements for the GPT prompt on cases where the landmark is the city name

1.3.2 - Bug fix on street name reverse geocoding

Update 4-AI_image_multimodal_location_GPT.R

1.3.1 - Improved reverse geocoding

Improvements on the reverse geocoding (street, city and country identification) algorithm, addressing issues in some specific situations where the address is not found (i.e. middle of a field or very far from any roads)

1.3.0 - Improved results from Bing Maps

Improved results from Bing Maps Location Recognition, now also searching for Business POI

1.2.1 - Bug fix for text generation

Correction of error that happened on the image description generation of files bigger than 4Mb

1.2.0 - Extra landmark identification source + Image Description

1. Adding Azure's Cognitive Services as an extra data source to recognize landmarks, adding to the similar Google Vision that was already being used. The script will use whatever of the 2 have a highest confidence score, since Azure doesn't provide GPS coordinates to check with image data;
2. Now also using Azure Vision API to generate a description of the image using AI.

1.1.0 - Minor improvements

Bug fixes on specific situations

1.0.0 - Initial Version

First release