Skin lesion detection AI

This invention

This invention is an AI-powered system that screens for melanoma and other skin lesions straight from smartphone-camera photos. A convolutional neural network, trained on dermatoscopic images, does the work. Inference runs on the device itself, so the system works offline. In plain terms, it puts a trained skin-lesion classifier in your pocket. It belongs to medical image analysis and computer-aided diagnosis — where deep learning, computer vision, and dermatology screening meet mobile hardware.

Where it fits

Your idea sits in a well-developed corner of Computer Vision (G06V) and AI & Machine Learning (G06N). Both run roughly 43× the corpus baseline here, a sign this is a tightly clustered, actively pursued space. It also draws on Image Processing (G06T) and Healthcare IT (G16H). Filings in this result set climbed steadily from 2016 onward, peaked around 2018–2020, and continued through 2024–2025. The area is genuinely active — recent years look lighter mainly due to publication lag, not a real slowdown. Groups working nearby in this set include 12 Sigma Technologies and Snap Inc (both prominent on citations), alongside established medical-imaging players. Being in such a lively neighborhood is normal, and it's a good sign your direction is real.

Closest related work

US-9886758-B2 — Annotation of skin image using learned feature representation (IBM · 5 citations · 2-member family)

This patent trains convolutional neural networks directly on dermatoscopic images, separating lesion skin from normal skin to annotate skin images. It tackles the same core problem you do — learning lesion features from dermatoscopic data. Reading it shows how IBM structured multiple CNNs around the lesion/normal boundary. It's a close conceptual neighbor for anyone building a skin-lesion classifier from dermatoscopic training data.

US-8543519-B2 — System and method for remote melanoma screening (Health Discovery Corporation · 33 citations · 2-member family)

This patent diagnoses skin conditions from images captured by a smartphone or digital camera and sent to an image-analysis server with a trained learning machine. It targets melanoma screening from consumer-camera photos — very close to your use case, though it relies on a server rather than on-device inference. Reading it shows how an early team framed the remote skin-screening pipeline end to end.

US-12239453-B2 — System and method for automatic personalized assessment of human body surface conditions (Little Angel Medical · filed 2025, recent · 0 citations · 2-member family)

This recent patent diagnoses skin, throat, and ear conditions from photographs taken on a mobile camera, using visual target overlays to reduce variability from camera pose. It's worth reading for how a current team handles the practical capture challenges — lighting, framing, and pose — that any smartphone-based skin screener must solve.

US-9414780-B2 — Dermoscopic data acquisition employing display illumination (Digimarc · 31 citations · 12-member family)

This patent uses a smartphone's own display to emit controlled spectral illumination while the camera gathers skin imagery from different angles. It addresses a complementary piece of your problem — getting consistent, dermatoscope-quality images from an ordinary phone. Reading it shows a clever hardware-light approach to improving input quality before any CNN ever sees the image.

What you can do next

• Explore & build on it. Browse the related work above — new, differentiated ideas often come from combining or improving on existing approaches, such as a specific on-device model architecture, a capture-standardization method, or an offline-inference optimization others haven't pinned down.
• If you'd like to protect it. Filing a provisional application (usually with a patent attorney) is a common first step. Most inventions can be protected in some form — what matters is how broad and defensible that protection is, which is where a patent attorney adds value (a very narrow claim may be granted but protect very little).
• If you'd like to make or sell it. The patents above point to who holds rights in this space; if your product would use a protected approach, licensing is a path worth exploring.