The 1.0 release of Perspec, a desktop application for correcting image perspective, has been announced. Developed over 9 years, Perspec aims to simplify the process of digitizing documents and receipts by allowing users to take photos and then refine them on their computers.

The developer expressed dissatisfaction with existing mobile scanner apps, citing issues like suboptimal file formats (e.g., JPEG for documents) and unhelpful live detection previews. Perspec prioritizes better image quality and a more efficient workflow for users who prefer desktop editing. A key technical improvement is its corner detection algorithm, which is argued to be more accurate for documents with non-straight edges than traditional edge detection methods used in apps like Adobe Scan and Dropbox.

The journey to 1.0 involved several technological iterations, starting with a Python CLI tool called Perspectra. The developer then moved to Haskell for a GUI, experimenting with ImageMagick and the Hip library before developing a pure C library, Flat CV, for computer vision and image manipulation. This allowed for the implementation of automatic corner detection directly within Perspec, fulfilling the vision for the 1.0 release.