o-mrost / microsoftapi

OOP II project with Microsoft Cognitive Services APIs

MicrosoftApi

OOP II project with Microsoft Cognitive Services APIs

Here will be our team project in which we use two of Microsoft Cognitive Services API: Computer vision and Bing image search. https://www.microsoft.com/cognitive-services/en-us/apis

The main idea of our app is to give user a possibility to upload their own picture, use url of an image or take image with webcam, analyse it with the API and based on the results look for similar images. The user will have possibility to fine tune the search parameters to include additional keywords or delete unnecessary ones. The idea is also to give user option to search for image of a certain size or particular license (for example for commercial use or for general purposes without modification).

• Computer vision API analyses image and returns tags and caption
• Bing Image Search searches for an image similar to the initial one based on tags and description and additional parameters included by the user.

To do list: https://docs.google.com/document/d/167Dvmo5O8xbwwDiLRfauzNLM1sfYpJbhUyaTTWPD0VE/edit?usp=sharing

microsoftapi's Issues

Webcam: null pointer exception

Rewrite the whole webcam thing

Analyse button is mixed up

When user

1. browses to image
2. analyses it
3. looks for similar images, and then
4. analyses the same image as before - different results are shown, and they seem to be the result of analysis of the last of four found images.

Search for similar images problem

App always finds less results than original Bing image search.
For example, with simple searches, such as "tree" or "ear" it shows only one image, and the rest are exceptions.
With queries such as "white car", or "red dress", or "blue water" it doesn't show any single image.

Why?
As an option - search for more than 4 images, then try-catch to show them on a label, if app goes to catch block because the image can’t be shown, null pointer exception or something similar, try to display the next image

Filechooser: name of file to be saved is not shown, it shows the name of previously opened file instead

When the user browsed a file, analysed it, searched for similar images and then tries to save it, they can choose the folder where the file will be saved, but the name of the file is shown incorrectly. Filechooser shows name of the file user browsed to, but saving works correctly, the file is saved with the name substringed from the url.
User also has no possibility to change the name of the file.