KM Research Hub

P1 · PANEL / ROI / AXIS GROUND TRUTH

原 EXE 对齐基准真值复核

所有保存均写入追加式审计事件;旧标注、旧比较和旧提取结果不会被删除或覆盖。

返回质量诊断
冻结记录120
真值完整117
待复核0
无效/退役3
← 上一例 Case 99 / 120 · DB #99 下一例 →
artifact_risk pending PMID 39542826

Biopsy-based Basal-luminal Subtyping Classifier in High-risk Prostate Cancer: A Combined Analysis of the NRG Oncology/RTOG 9202, 9413, and 9902 Phase 3 Trials.

F2 · Panel Luminal_Subtypes · Prostate Cancer Specific Mortality

Fig. 2 – (A) Cumulative incidence of prostate cancer–specific mortality (PCSM) by duration of androgen deprivation therapy (long-term [≥24 mo] vs short-term [4 mo] ADT) and prostate subtyping classifier (PSC) category (basal or luminal), with adjusted subdistributional hazard ratios (sHRs) and their 95% confidence intervals (CIs) estimated from multivariable Fine and Gray interaction models. (B) Differences in restricted mean survival times (in months with corresponding 95% CIs) by duration of ADT (long term [≥24 mo] vs short term [ < 4 mo]) for basal and luminal subtypes for all endpoints at 5-, 10-, and 15-yr follow-up. Note that basal subtype did not have 15-yr follow-up for the biochemical failure endpoint. ADT = androgen deprivation therapy; CI = confidence interval; ΔRMST = difference in restricted mean survival time; sHR = subdistributional hazard ratio.
打开论文记录

坐标框复核

洋红 = 当前 Linux 候选;绿色 = 已保存人工真值。没有绿色框表示仍未形成真值。

Case 99 ROI review overlay
当前候选 {'x0': 118, 'y0': 7, 'x1': 843, 'y1': 289} 人工真值 {'x0': 118, 'x1': 843, 'y0': 7, 'y1': 289}
生产预期曲线2
当前恢复曲线1
算法版本legacy-0.5.0
坐标 OCR完整
查看候选 OCR 与警告(只能作种子,不能自证真值)
{
  "complete": true,
  "engine": "RapidOCR",
  "plot_bottom": 288,
  "plot_left": 118,
  "plot_right": 842,
  "plot_top": 7,
  "warnings": [],
  "x_axis_y": 304,
  "x_confidence": 0.832,
  "x_detected": true,
  "x_max": 22,
  "x_min": -1,
  "x_ticks": [
    {
      "box": [
        [
          204.0,
          309.0
        ],
        [
          221.0,
          309.0
        ],
        [
          221.0,
          325.0
        ],
        [
          204.0,
          325.0
        ]
      ],
      "confidence": 0.9992951154708862,
      "pixel": 212.5,
      "text": "2",
      "value": 2.0
    },
    {
      "box": [
        [
          237.0,
          310.0
        ],
        [
          250.0,
          310.0
        ],
        [
          250.0,
          323.0
        ],
        [
          237.0,
          323.0
        ]
      ],
      "confidence": 0.99809330701828,
      "pixel": 243.5,
      "text": "3",
      "value": 3.0
    },
    {
      "box": [
        [
          270.0,
          311.0
        ],
        [
          281.0,
          311.0
        ],
        [
          281.0,
          322.0
        ],
        [
          270.0,
          322.0
        ]
      ],
      "confidence": 0.9993380904197693,
      "pixel": 275.5,
      "text": "4",
      "value": 4.0
    }
  ],
  "y_axis_x": 115,
  "y_confidence": 0.912,
  "y_detected": true,
  "y_max": 100,
  "y_min": 0,
  "y_ticks": [
    {
      "box": [
        [
          83.0,
          65.0
        ],
        [
          111.0,
          65.0
        ],
        [
          111.0,
          89.0
        ],
        [
          83.0,
          89.0
        ]
      ],
      "confidence": 0.9999138712882996,
      "pixel": 77.0,
      "text": "75",
      "value": 75.0
    },
    {
      "box": [
        [
          84.0,
          135.0
        ],
        [
          111.0,
          135.0
        ],
        [
          111.0,
          160.0
        ],
        [
          84.0,
          160.0
        ]
      ],
      "confidence": 0.9997358620166779,
      "pixel": 147.5,
      "text": "50",
      "value": 50.0
    },
    {
      "box": [
        [
          84.0,
          205.0
        ],
        [
          111.0,
          205.0
        ],
        [
          111.0,
          230.0
        ],
        [
          84.0,
          230.0
        ]
      ],
      "confidence": 0.999823659658432,
      "pixel": 217.5,
      "text": "25",
      "value": 25.0
    }
  ]
}
查看当前 Linux 输出

追加一条审查结论

选择“有效 KM”时,四边 ROI、四个坐标范围、Panel、终点和曲线数均为必填。

像素 ROI(x1/y1 为排他边界)
真实坐标范围

追加式审计历史

最新事件仅作为当前缓存;以下每一条既往结论都永久保留。

事件时间 / 审查人状态ROI坐标说明
#102p1-visual-axis-audit-v2 · frozen benchmark contact sheets + original images + RapidOCR re-audit 2026-07-28T15:12:18+00:00Codex-assisted visual audit valid_km {'x0': 118, 'x1': 843, 'y0': 7, 'y1': 289} X 0.0–21.0 / Y 0.0–100.0 P1 冻结原图审查;人工检查冻结原图刻度;ROI 四边按可见坐标轴框记录。

最近一次正式比较

Suite #3 · 2026-07-28T15:15:12+00:00

未通过

真值完整 117 / 120;通过 0;缺真值 0;无效案例 3。