Conversation
WebARKitOrbTracker classes. Now only WebARKitTracker is necessary. - fix for issue webarkit/webarkit-testing#11
Feature webarkit trackers stable
- FREAK and TEBLID
- new WebARKitManager class to be improved
Add an opt-in option so the tracked pose origin is the marker centre instead of the reference image's top-left corner -- AR content at (0,0,0) then sits in the middle of the marker. Default OFF (matches ArtoolkitX, which never centres at any layer, and preserves existing behaviour). When enabled, the 3D object points are offset by (_pattern.size/2) just before solvePnP: this moves only the solved translation, leaves the rotation unchanged, and the 2D matching/template/homography paths untouched. Exposed as setOriginCentered(bool) via the tracker and manager. Refs webarkit/webarkit-testing#38 Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Restore ArtoolkitX-style detection-side pyramid downsampling: feature detection now runs on a pyrDown'd copy of the live frame (the "detectionFrame"), and matched frame keypoints are scaled back to full-frame coordinates before the homography fit. - initialize(): compute _featureDetectPyrLevel from featureImageMinSize (clamped >= 0) and the exact cv::pyrDown scale factor via the iterative (x+1)/2 rounding, so the keypoint rescale is sub-pixel accurate. - processFrame(): build detectionFrame via _featureDetectPyrLevel pyrDown steps; run createFeatureMask/extractFeatures on it. - MatchFeatures(): rescale matched frame keypoints by _featureDetectScaleFactor before getHomographyInliers. - createFeatureMask(): divide the warped bbox by the scale factor so the tracked-marker exclusion lands in downsampled coordinates. Level 0 (frame <= featureImageMinSize, e.g. 640x480) => detectionFrame == frame and factor 1.0, so that path is byte-identical to full-res detection. The reference image stays detected at full resolution (initTracker), only the live frame is downsampled, matching ArtoolkitX OCVT. This is part A of #44. The "skip detection while tracking" guard is a separate follow-up (part B). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Guard the per-frame feature detection (extractFeatures + descriptor matching) with `if (!_isTracking)`. Once optical flow holds a lock, detection is skipped and the pose is maintained by optical flow + template matching; when the marker is lost (_isTracking cleared in runOpticalFlow / RunTemplateMatching), detection resumes on the next frame to re-acquire. Detection is the dominant per-frame cost, so skipping it in the steady tracking state is a large win: at 640x480 the webcam example goes from ~10-15 fps (detecting every frame) to ~45 fps while tracking. This is part B of #44; part A (detection-side pyramid downsampling) landed in #52. The guard is on _isTracking rather than ArtoolkitX's `_currentlyTrackedMarkers < _maxNumberOfMarkersToTrack`: _isDetected is reset to false every frame, and optical flow is skipped on the first detection frame (_frameCount == 0), so a counter-based guard would skip both detection and optical flow on the following frame and freeze (this bites the static-image example, which feeds the same frame repeatedly). Gating on _isTracking re-detects until optical flow actually holds a lock. The optical-flow/template/pose paths (gated on _isDetected || _isTracking) and #46 tracking-loss are unchanged. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
getPoseMatrix() returned a zero matrix: it reads _patternTrackingInfo.pose3d, which the live tracking path never wrote (the pose flows _pose -> transMat -> trans; pose3d was touched only by the dead computePose). cameraPoseFromPoints now also builds pose3d (the raw 4x4 OpenCV-convention pose) from the same rMat/tvec, so the CV getter reflects the current frame. Rename the two pose getters so the distinction is explicit (they differ in both convention and shape, which the "2" suffix hid): getPoseMatrix -> getPoseMatrixCV (raw OpenCV 4x4 pose) getPoseMatrix2 -> getPoseMatrixGL (right-handed / GL-corrected pose, trans) Hard rename across WebARKitManager + WebARKitTracker (decls, inner impl, outer wrapper). The matching emscripten bindings / controller live in webarkit-testing. Verified: static (1920x1440) and webcam (640x480) examples both track (1000 matches); getPoseMatrixCV() now returns a valid [R|t;0001] (was all zeros). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
update() was declared in the header but never defined or called (the live per-frame path is processFrameData()). It only linked because nothing referenced it; calling it would have been a link error. Not scaffolding -- a future update() would be declared with its implementation. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
im_gray() was a fully commented-out grayscale routine in WebARKitUtils.h, superseded by convert2Grayscale(). Pure dead text (git history retains it). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Remove the commented-out multi-marker _trackables[...] lines (the old artoolkitX-style forms left beside their live single-marker replacements) and the commented getPoseMatrix3 declarations. Comment-only; the live code and explanatory comments are untouched, so the compiled output is identical. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
…Matrix() These were WebARKit's own pose-math path, superseded by cameraPoseFromPoints + the #55 pose3d fix: - computePose() -- its only call site was already commented out; cameraPoseFromPoints is the live pose solve. - invertPose() -- only ever called by computePose(). - computeGLviewMatrix() (no-arg) -- uncalled; the live path uses the computeGLviewMatrix(cv::Mat&) overload (kept). Also drops the dangling commented computePose call in WebARKitTracker.cpp. Behavior-neutral; the live pose/GL-view path (cameraPoseFromPoints, getTrackablePose, updateTrackable, computeGLviewMatrix(cv::Mat&)) is unchanged. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
README: replace the stub with a developer-oriented map — the two lineages (WebARKit/ = ArtoolkitX-OCVT-derived active code vs vendored lib/SRC + include/AR = old ARToolKit5, built by jsartoolkitNFT; ARnft consumes jsartoolkitNFT as an npm module), code layout, pose pipeline, who-builds-what (webarkit-testing builds WebARKit/ to WASM via makem.js, CMake migration planned), the gtest/CMake tests, and the contribution path. version: 0.8.0 (was 1.0.0) -- 1.0.0 should denote a complete/LTS release, which this isn't yet. Updates WEBARKIT_HEADER_VERSION_STRING + MAJOR/MINOR. Refs #51. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The version bump (1.0.0 -> 0.8.0) broke the gtest version checks: WEBARKIT_HEADER_VERSION_STRING and getWebARKitVersion() are now "0.8.0". Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Note in the Building section that WebARKit/CMakeLists.txt can already build WebARKitLib as a static library today; the "planned" migration is specifically webarkit-testing moving its WASM build off tools/makem.js onto it. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
…dency Clarify the CMake static-lib build: it targets both WASM (emscripten) and native Linux via EMSCRIPTEN_COMP, and depends on a prebuilt OpenCV from webarkit/opencv-em (emscripten opencv-js build for WASM, native build for Linux), fetched via FetchContent. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Contributing section now requires commit messages to follow Conventional Commits and links to ./CONTRIBUTING.md (to be added) for the full guidelines. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add contribution guidelines: branch-from-dev / PR-to-dev, signed commits, Conventional Commits (with examples), what may/may not change (WebARKit/ vs the vendored old-ARToolKit5 lib/SRC), the cross-repo PR-pair flow, and build/test (CMake static lib for WASM/Linux via opencv-em; gtest). README already links here. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Note that the cross-repo PR-pair flow applies to WebARKit/ folder changes consumed by the webarkit-testing repo as WASM; vendored lib/SRC + include/AR (old ARToolKit5) changes instead flow through jsartoolkitNFT and don't use it. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
kalwalt
added
documentation
kalwalt
moved this from Review in progress
to Reviewer approved
in New markerless image tracking
github-project-automation
Bot
moved this from Reviewer approved
to Done
in New markerless image tracking
kalwalt
added a commit
to webarkit/webarkit-testing
that referenced
this pull request
Jun 21, 2026
Point the submodule at WebARKitLib's master tip (the dev->master integration, webarkit/WebARKitLib#60) instead of dev, and rebuild build/ + dist/. master now contains the full OCVT tracker line (== dev) plus the vendored AR/NFT side. Verified: rebuilt cleanly; static + webcam Teblid examples track (v0.8.0). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
kalwalt
added a commit
to webarkit/webarkit-testing
that referenced
this pull request
Jun 21, 2026
Point the submodule at WebARKitLib's master tip (the dev->master integration, webarkit/WebARKitLib#60) instead of dev, and rebuild build/ + dist/. master now contains the full OCVT tracker line (== dev) plus the vendored AR/NFT side. Verified: rebuilt cleanly; static + webcam Teblid examples track (v0.8.0). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Integration of the
devbranch (the WebARKit OCVT planar-tracker line) intomaster. 167 commits ahead of the merge base;masterhas 13 of its own (the NFT/AR vendored side). The two lines touch disjoint subsystems, so this is a clean integration rather than a conflict-heavy merge.Merge mechanics (verified)
git merge-treereports zero conflicts — the textual merge is clean.dev=WebARKit/(the OCVT tracker);master= vendored AR / NFT / KPM / luma (lib/SRC,include/AR).devtouches nothing underlib/SRC/include/AR.include/AR/config.h→1.7.6(ARToolKit5 / NFT, consumed by jsartoolkitNFT) — frommaster.WebARKit/.../WebARKitConfig.cpp→0.8.0(the OCVT tracker) — fromdev.What
devbrings (the OCVT tracker,WebARKit/)Pose / projection / handedness
getPoseMatrixzero-matrix fix + rename togetPoseMatrixCV(raw OpenCV 4×4) /getPoseMatrixGL(right-handed) (getPoseMatrix() returns a zero matrix — pose3d is never populated in the live path #55).setOriginCentered(Fix crash when a marker is detected on the first frame (#37) #38).Tracking robustness & performance
Examples
Code health & docs
update()decl,im_gray(), commented multi-marker scaffolding, and the supersededcomputePose/invertPose/computeGLviewMatrix()-noarg; kept (with rationale) ArtoolkitX-parity symbols and debug scaffolding (Code audit: find duplicated / orphan / unused code in WebARKit (excluding vendored lib/SRC and AR/ headers) #50).CONTRIBUTING.md(Conventional Commits, cross-repo PR-pair flow) + version0.8.0(README / developer docs: explain WebARKit code structure, build flow, and contribution path #51).Validation
test.ymlgtest) runs on this PR — covers the OCVT side (already green ondev).master's vendored AR/luma/KPM changes) + a webarkit-testing submodule bump tomaster+ rebuild/examples (OCVT WASM).🤖 Generated with Claude Code