KISS-Matcher about: drop the single-Vel64-sample sentence (now multi-preset)

Author: LimHyungTae

web/src/i18n/locales/en.ts

@@ -389,7 +389,7 @@ export const en = {
       title: "KISS-Matcher — Global Registration",
       subtitle: "No initial guess needed: FasterPFH descriptors → ROBIN matching → GNC solver.",
       about:
-        "KISS-Matcher (Lim et al., ICRA, 2025) is a fast and robust global point cloud registration pipeline. It chains FasterPFH (a faster Point Feature Histogram descriptor), ROBIN graph-based outlier matching, and a Graduated Non-Convexity (GNC) solver to recover SE(3) without any initial pose guess. This demo replays a precomputed run on the official Vel64 sample (KITTI HDL-64 frames 540 / 1319 shipped with the KISS-Matcher repo); slate lines are the raw initial correspondences after cross-checking, amber lines are the final inliers that survived ROBIN + GNC, and \"Apply T\" applies KISS-Matcher's recovered transform to the source cloud.",
+        "KISS-Matcher (Lim et al., ICRA, 2025) is a fast and robust global point cloud registration pipeline. It chains FasterPFH (a faster Point Feature Histogram descriptor), ROBIN graph-based outlier matching, and a Graduated Non-Convexity (GNC) solver to recover SE(3) without any initial pose guess. Slate lines are the raw initial correspondences after cross-checking, amber lines are the final inliers that survived ROBIN + GNC, and \"Apply T\" applies KISS-Matcher's recovered transform to the source cloud.",
       params: [
         { name: "voxel_size", desc: "Set in tools/gen_kiss_matcher_data.py and baked into the precomputed JSON. KISS-Matcher's only knob — keypoint density and correspondence count both follow from it.", effect: "" },
         { name: "Original / Apply T", desc: "Toggle between the unaligned src and KISS-Matcher's recovered alignment.", effect: "" },

web/src/i18n/locales/ko.ts

@@ -383,7 +383,7 @@ export const ko: LocaleDict = {
       title: "KISS-Matcher — Global Registration",
       subtitle: "초기 추정값 없이도 동작: FasterPFH 디스크립터 → ROBIN matching → GNC solver.",
       about:
-        "KISS-Matcher (Lim et al., ICRA, 2025)는 빠르고 강건한 글로벌 point cloud registration 파이프라인입니다. FasterPFH(빠른 Point Feature Histogram 디스크립터), ROBIN graph 기반 outlier matching, Graduated Non-Convexity (GNC) solver를 차례로 적용해 어떠한 초기 pose 추정값 없이도 SE(3)을 복원합니다. 본 데모는 KISS-Matcher 공식 저장소에 포함된 Vel64 샘플(KITTI HDL-64 540 / 1319 프레임) 쌍에 미리 돌린 결과를 그대로 보여줍니다. slate 색 선은 cross-check 직후의 초기 correspondence, amber 색 선은 ROBIN + GNC를 거쳐 살아남은 최종 inlier이며, \"Apply T\"는 KISS-Matcher가 추정한 transform을 src에 적용합니다.",
+        "KISS-Matcher (Lim et al., ICRA, 2025)는 빠르고 강건한 글로벌 point cloud registration 파이프라인입니다. FasterPFH(빠른 Point Feature Histogram 디스크립터), ROBIN graph 기반 outlier matching, Graduated Non-Convexity (GNC) solver를 차례로 적용해 어떠한 초기 pose 추정값 없이도 SE(3)을 복원합니다. slate 색 선은 cross-check 직후의 초기 correspondence, amber 색 선은 ROBIN + GNC를 거쳐 살아남은 최종 inlier이며, \"Apply T\"는 KISS-Matcher가 추정한 transform을 src에 적용합니다.",
       params: [
         { name: "voxel_size", desc: "tools/gen_kiss_matcher_data.py에서 설정하고 precomputed JSON에 그대로 박힙니다. KISS-Matcher의 유일한 knob, keypoint 밀도와 correspondence 개수가 모두 여기서 결정됩니다.", effect: "" },
         { name: "Original / Apply T", desc: "정렬되지 않은 src와 KISS-Matcher가 정렬한 src 사이를 토글.", effect: "" },