The article discusses the challenges of comparing images in visual test automation, emphasizing that while AI excels at recognizing image content, it struggles with structural differences. Conventional methods require precise pixel-level alignment, leading to issues with minor distortions. Convolutional neural networks (CNNs) are suggested as a solution, working better by segmenting images into smaller regions. However, high-resolution displays could result in false positives by identifying minor displacements as significant changes. The article underscores the need for improved detection methods to achieve more reliable visual testing results.
Spotting differences between two images is a relevant task in visual test automation when a screenshot needs to be compared to a previous version or design.
While multimodal language models excel at recognizing and explaining content, they struggle with detecting differences unless explicitly trained on those aspects.
Collection
[
|
...
]