{"slug":"machine-vision","title":{"ja":"マシンビジョン","en":"Machine vision"},"summary":{"ja":"機械が見ている対象を、検出枠・ラベル・偽色の解析画面で表す映像の様式。","en":"The world shown the way a computer perceives it: detection boxes, labels and false-color analysis passes."},"categories":["motion","uiux","3dcg"],"subAspects":["motion.style"],"aspects":["motion"],"tags":["machine vision","computer vision","object detection","segmentation","ai","surveillance"],"palette":[{"name":"Footage dark","hex":"#0B0F14","ratio":0.55},{"name":"Street gray","hex":"#3A4250","ratio":0.2},{"name":"Detection green","hex":"#39FF88","ratio":0.12},{"name":"Cyan","hex":"#4FD8FF","ratio":0.08},{"name":"Lock red","hex":"#FF3B3B","ratio":0.05}],"keywords":["machine vision","computer vision overlay","object detection bounding boxes","semantic segmentation","depth map","edge detection","pose estimation skeleton"],"traits":{"en":["Thin 1–2px rectangular boxes surround each detected object, with corner brackets and a small filled label tag reading the class in uppercase monospace with a confidence value (PERSON 0.97)","Every box and line is one saturated hue (green or cyan); a single locked target uses a contrasting hue (red or amber) with a crosshair","False-color analysis passes replace the picture: flat segmentation colors per class, a near-bright depth ramp, white edge-line drawing or a heatmap of attention","People are drawn as keypoint skeletons (dots joined by colored limb lines), and motion as a grid of small arrows","Passes wipe in and out along a straight divider; a scan line sweeps at a constant rate and boxes flicker briefly when acquired"],"ja":["検出した物それぞれを、1〜2pxの細い四角い枠と四隅の括弧で囲み、小さな塗りのラベルに種類を大文字の等幅で、信頼度と並べて書く（PERSON 0.97）","枠と線はすべて彩度の高い1色（緑かシアン）。1つだけロックした対象は対比の色（赤か琥珀）で、照準を重ねる","偽色の解析画面が映像に置き換わる。種類ごとの平塗り（セグメンテーション）、近いほど明るい深度、白い輪郭の線画、注目度のヒートマップ","人は関節の点を色のついた線でつないだ骨格で、動きは小さな矢印の格子で表す","解析画面は直線の境目でワイプして切り替わる。走査線が一定の速さで動き、枠は捉えた瞬間に短くちらつく"]},"prompt":"Machine-vision video overlay: a dark street scene seen by a computer, every person and vehicle enclosed in thin 1–2px green bounding boxes with corner brackets and small filled label tags in uppercase monospace with confidence values (PERSON 0.97, VEHICLE 0.88), one target locked in red with a crosshair. Half of the frame is wiped to a false-color semantic segmentation pass with flat colors per class and white edge lines. A horizontal scan line, keypoint skeletons on pedestrians, minimal glow.","negativePrompt":"sci-fi cockpit HUD, circular gauges, robot eye, iris scan, real faces, cluttered full-screen interface, warm cinematic grading","contrasts":[],"images":[],"references":[{"label":"remotion-video-ui-library — Effects/Overlay/ObjectDetectionOverlay","url":"https://github.com/asamiile/remotion-video-ui-library/tree/main/src/Effects/Overlay/ObjectDetectionOverlay"},{"label":"remotion-video-ui-library — Background/SceneUnderstanding","url":"https://github.com/asamiile/remotion-video-ui-library/tree/main/src/Background/SceneUnderstanding"},{"label":"remotion-video-ui-library — Effects/Stylize/FootagePass","url":"https://github.com/asamiile/remotion-video-ui-library/tree/main/src/Effects/Stylize/FootagePass"},{"label":"remotion-video-ui-library — Background/PoseEstimation","url":"https://github.com/asamiile/remotion-video-ui-library/tree/main/src/Background/PoseEstimation"},{"label":"remotion-video-ui-library — Background/OpticalFlow","url":"https://github.com/asamiile/remotion-video-ui-library/tree/main/src/Background/OpticalFlow"}],"links":{"page":{"ja":"https://visual-gallery.asami.tokyo/ja/styles/machine-vision","en":"https://visual-gallery.asami.tokyo/en/styles/machine-vision"},"json":"https://visual-gallery.asami.tokyo/api/v1/styles/machine-vision","markdown":"https://visual-gallery.asami.tokyo/api/v1/styles/machine-vision/md"}}