← Help center

GALLERY: REVIEWING A DETECTION

RGB / DEPTH / 3D on one bar, with edit tools and Save view

OPENING A DETECTION

Open the gallery (G), then click a detection card to open the fullscreen viewer.

REVIEWING EVIDENCE FROM ANOTHER FOLDER

You can review detections that were captured on a different machine. Copy the detections folder to your computer, then point SPECTER at it:

  1. Open Settings and go to the Recording tab.
  2. Under DETECTION SAVE FOLDER, click Choose and select the folder that contains the session_* subfolders.
  3. Open the gallery. Every detection in that folder now appears, grouped by session.

Keep each detection's files together and unrenamed. The gallery reads them by name, and depth captures need their depth files intact to rebuild the 3D view. For best performance with 3D depth, review from a local or fast external drive rather than a slow network share.

BOTTOM LAYER BAR

ButtonAction
RGBAnnotated still (detection box baked onto the image).
DEPTHDepth map re-rendered from raw sensor data (editable).
3D ORBITLive orbitable point cloud rebuilt from the raw depth map. Drag to rotate, scroll to zoom. The detection is rendered inside the reconstructed room, with its classification label, so you can see where in space it registered.
3D SNAPBaked static 3D render, shown when the capture stored a rendered cloud but no raw depth to orbit.

Keyboard: D toggles RGB ↔ DEPTH (not 3D).

Captures made without a depth sensor active show NO DEPTH DATA in place of the 3D layer. The visual layers stay available.

THUMBNAIL SOURCE

The detection rail and gallery grid can show each hit as any of its saved channels. Use the thumbnail source selector to switch every thumbnail between RGB, DEPTH (heatmap), DEPTH GREY, THERMAL, or IP CAMERA. Grey depth thumbnails are rendered live from the raw depth data and honor the same GAIN setting, so older detections captured before a channel existed simply fall back to RGB. Your choice is remembered between sessions.

EDIT PANEL (✎ EDIT)

Image tools (RGB and DEPTH)

Depth tools (visible when the DEPTH layer is active)

On the DEPTH layer, the red detection box is drawn from saved metadata onto the rendered depth map.

SAVING

ButtonOutput
Save viewgallery_view_NN.png + gallery_view_NN.json sidecar (layer, filters, range, color mode, bbox info).
Save as new versionsnapshot_edited_NN.png, a classic flattened PNG export.

WHAT GETS SAVED WITH A DETECTION

On a depth-capable sensor, each auto-detection folder may contain:

FileDescription
snapshot.pngClean RGB plate
snapshot_annotated.pngRGB + overlays + red detection box
snapshot_depth.png2D depth heatmap at capture time
depth_raw.binRaw uint16 depth grid
depth_rgb.jpgRGB image registered to depth resolution
depth_intrinsics.jsonCamera geometry for 3D rebuild
snapshot_3d.pngBaked 3D render (3D SNAP layer)
metadata.jsonScores, bbox, intensity class, global timecode, mode, flags
clip.mp4Pre-event video buffer
gallery_view_*.pngOptional user exports from the gallery
Badge legend on gallery cards: D = a 2D depth heatmap was saved; 3D = an orbitable 3D depth snapshot (depth_raw.bin) was saved.

RELATED