Company name
Github
Problem Statement
Markdown alt text suggestor - GitHub Action that can be enabled by a maintainers to scan markdown documents looking for inline images missing alt text and suggesting them. This can be run as part of a PR pre-merge check or periodically on the codebase to help improve the accessibility of a project documentation set.
What it does
This Flask app is designed to help improve the accessibility of a project's documentation set by suggesting alt text for inline images that are missing them. When a user enters a URL of a web page that contains inline images, the app processes the page and generates markdown suggestions that can be added to the alt text attribute of each image. The suggestions are based on computer vision algorithms provided by Microsoft Azure Cognitive Services, which analyze the content of each image to generate relevant alt text descriptions. The app then displays the markdown suggestions in a new page, making it easy for the user to copy and paste them into their documentation. Overall, this app can help project maintainers ensure that their documentation is accessible to everyone, including those who may rely on assistive technologies to navigate the web.
How we built it
The project is built using a combination of technologies, including Python, Flask, and Microsoft Azure Cognitive Services.
The Python language is used to write the core functionality of the app, including the code to scrape the web page, extract the images, and generate the alt text suggestions using Azure Cognitive Services.
Flask, a popular Python web framework, is used to create the web app itself, including the user interface and the logic to handle incoming requests and display the output.
Finally, Microsoft Azure Cognitive Services is used to provide the computer vision and natural language processing capabilities needed to analyze the images and generate the alt text suggestions. Specifically, the Computer Vision API is used to identify objects within the images, while the Text Analytics API is used to generate natural language descriptions of those objects.
Challenges we ran into
One of the challenges that could arise in this app is if the app is not generating the expected alt text suggestions for the images. There could be several reasons for this issue, including:
Insufficient or low-quality images: The accuracy of the alt text suggestions depends on the quality and resolution of the images. If the images are of low quality or resolution, it may be difficult for the computer vision algorithms to accurately identify the objects in the images and generate the correct descriptions.
Language or cultural barriers: The Text Analytics API used in this app to generate natural language descriptions may not be able to accurately understand or describe certain objects, depending on the language or cultural context in which they appear.
Incorrect image detection: The Computer Vision API used in this app to identify objects in the images may not be able to accurately detect all objects or may misidentify objects, leading to inaccurate or incomplete alt text suggestions.
Technical issues: Technical issues such as connectivity problems, API errors or compatibility issues between the app and the APIs may also impact the alt text suggestions generated by the app.
Log in or sign up for Devpost to join the conversation.