@yefoi

so its like the Automator app in mac but its more user friendly, it looks better, and it uses scratch style blocks to drag and drop around instructions and stuff like "wait 1 second", "move mouse to [x], [y]", "loop [x]" and stuff like that Automator for Mac right now is ugly and unintuitive primarily just gonna be a macro maker for keys and mouse presses but i'll look into being able to access browser pages and stuff

The purpose of this paper is to see whether or not a custom algorithm that relies purely on color images can identify vegetation health, how accurately it can do so, how resource efficient such an algorithm is, and how it compares to a YOLOv8 model specifically fine-tuned to identify and categorize these images. The way that the algorithm works is by essentially creating a color histogram of each pixel in the image, and comparing it to a list of reference images that the user categorizes themselves. The distribution that is the closest are the one that gets categorized, and the more distinctly similar the input distribution is to one of the reference distributions, the more “confident” the model is. If the user provides multiple reference images, then the reference for each category is averaged. In addition, the texture of each image is also measured in terms of how significant shifts in value are in HSV, symbolizing large spots of shadows and highlights.

Alternative app for Roadready, government endorsed driving logging app for teens like me, yet simultaneously hasn't been updated since 2015 (It shows). Being made in react native for cross compatibility.