centersl/MinerU
1
0
Fork 0
mirror of https://github.com/opendatalab/MinerU.git synced 2026-03-27 11:08:32 +07:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #1792 from opendatalab/dev

Browse Source 
fix: match multiple captions
This commit is contained in:
Xiaomeng Zhao
2025-02-27 10:48:42 +08:00
committed by GitHub
parent b6d60bb0fa 2629cadca6
commit 97633eb1e2
1 changed files with 36 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

36
magic_pdf/model/magic_model.py
View File

@@ -553,6 +553,42 @@ class MagicModel:
}
)
for i in range(len(objects)):
j = i + OBJ_IDX_OFFSET
if j in seen_idx:
continue
seen_idx.add(j)
nearest_dis, nearest_sub_idx = float('inf'), -1
for k in range(len(subjects)):
dis = bbox_distance(objects[i]['bbox'], subjects[k]['bbox'])
if dis < nearest_dis:
nearest_dis = dis
nearest_sub_idx = k
for k in range(len(subjects)):
if k != nearest_sub_idx: continue
if k in seen_sub_idx:
for kk in range(len(ret)):
if ret[kk]['sub_idx'] == k:
ret[kk]['obj_bboxes'].append({'score': objects[i]['score'], 'bbox': objects[i]['bbox']})
break
else:
ret.append(
{
'sub_bbox': {
'bbox': subjects[k]['bbox'],
'score': subjects[k]['score'],
},
'obj_bboxes': [
{'score': objects[i]['score'], 'bbox': objects[i]['bbox']}
],
'sub_idx': k,
}
)
seen_sub_idx.add(k)
seen_idx.add(k)
for i in range(len(subjects)):
if i in seen_sub_idx:
continue