TABLE 1.
Category | # | Feature | Definition | Example |
---|---|---|---|---|
Metadata | 1 | Free text/structured fields | Option to provide additional information about an image prior to submission, either through free text fields or writing in/selecting options in structured fields |
Free text: After uploading an image of their skin, the user can write in additional comments in a free text box prior to submission. Structured fields: User is asked to answer pre‐written, multiple‐choice questions about the duration, morphology, and evolution of their skin lesion prior to image submission. |
2 | Timestamp/temporal labeling | Automatic labeling and/or sorting of images by the date and time uploaded or submitted | When navigating to previously uploaded images, the user sees them arranged in date‐time order with grey timestamps beneath each thumbnail. | |
3 | 2D body map tracking | Ability to identify anatomic location of skin complaint in two‐dimensional space, usually on an avatar | After taking an image of their skin, the user sees a 2D anatomical model and is asked to tap the location closest to that of the real‐life skin lesion. | |
4 | Image/album labeling | Ability to retitle images or groups of images organized into folders/albums | User selects 3 photos out of the 15 that they have uploaded and moves them to a new in‐app folder, which they title “Rash R Arm March 2020.” | |
5 | Lesion measurement | Automated (in‐camera or device‐based measurement scale/frame) or guided (with reference object or user‐based estimate) measurement of lesion size | User is instructed to include and identify a reference object (quarter) in their skin image. After the image is uploaded, “2 × 3mm” appears automatically in the size field. | |
6 | 3D body map tracking | Ability to identify anatomic location of skin complaint in three‐dimensional space, usually on an avatar | After taking an image of their skin, the user sees a 3D anatomical model and is asked to rotate, zoom, and then tap on the location closest to that of the real‐life skin lesion. | |
Functional tools | 7 | AI diagnostic analysis | Application returns some type of diagnostic output with respect to an uploaded image; outputs can be binary, risk class‐based, a list of ranked or unranked diagnoses, and/or a score on a continuous risk scale | After submitting an image to the “AI Analyzer,” the user sees results suggesting that their lesion has a 57.1% chance of malignancy and that they should contact a dermatologist. |
8 | Digital share | Ability to share labeled images/image albums, lesion analyses, symptom visualizations, or other in‐app media via email or through other smartphone apps (i.e., text messaging) | When viewing a graph tracking the severity of itch associated with uploaded images over time, the user sees a “Share” button that automatically attaches the PDF to a new email. | |
9 | Photo reminders | Ability to set and/or customize in‐app reminders to take skin images | A psoriasis patient receives weekly reminders every Saturday to take and upload images of their affected skin. | |
10 | Compare lesions | Ability to see two or more skin images side by side | After selecting two previously uploaded images and clicking “Compare,” the user sees one on the left side of the screen and the other on the right. | |
11 | In‐app zoom | Ability to zoom in on or magnify any area of an uploaded image | User selects a photo and pinches in with two fingers to get a magnified view of any area. | |
12 | File format conversion |
In‐app conversion of uploaded skin images between different file types |
After uploading a DICOM file to the app interface, the user can choose to download it as a JPG, PNG, or TIFF from within the same application. | |
Image processing | 13 | Image cropping | Removal of unwanted outer areas from uploaded skin image | After uploading an image, the user trims/removes outer areas representing extraneous background by moving the image borders closer together. |
14 | Magnification to resize | Ability to resave a magnified/zoomed in version of a previously uploaded image | After uploading an image, the user pinches in with two fingers to magnify the lesion of concern and saves the magnified version of the image. | |
15 | Image realignment/orientation | Ability to rotate a previously uploaded image, most commonly to reorient it with a previous image or to facilitate ease of interpretation | After uploading an image, the user hits an arrow button to rotate the image 90°. | |
16 | Color/contrast correction | Automated or manual ability to adjust color contrast, saturation, filters, and so on, usually with the goal of more accurately representing real‐life appearance | After noticing that their uploaded photo has an orange cast due to harsh lighting, the user adjusts the image contrast and yellow/orange tones using an in‐app color filter. | |
17 | Image markup | Ability to add text boxes, digital pen markings, or other artificial elements to the image | After uploading an image, the user adds an arrow pointing to the lesion of interest. | |
18 | Light balance correction | Automated or manual ability to adjust light balance, usually with the goal of more accurately representing real‐life appearance | After taking an image in a shadowy room, the user uses the light adjustment feature to “brighten” the image so that the lesion of interest is more visible. | |
19 | Obscuration removal | Automated or manual removal of anything covering the lesion/area of interest | After noticing that their bracelet partially obscures their rash in a previously uploaded skin image, the user presses “Remove Obscuration” to replace it with a skin‐colored area. | |
20 | Background removal | Automated or manual removal of extraneous, non‐skin image background | After noticing a brightly colored shirt in the background of a previously uploaded skin image, the user presses “Remove Background” to replace all non‐skin background with solid grey. |