9i4SSKJr SSKrSSKJr SSKJrJr SSK r SSK J r SSKJr SSKJr SSKJrJr SSKJr SSKJrJr SS KJr \R8"\\R:S 9"S S \55rS rg)N) OrderedDict)AnyDict)COCO)COCOeval)dumpload)Metrics)METRICSMetric) default_group) group_key module_namec\rSrSrSrSrS\\\4S\\\44Sjr Sr SSjr S r S r SS jrS rSrSrSrSrSrg )#ImageInstanceSegmentationCOCOMetriczMThe metric computation class for COCO-style image instance segmentation. c SUlSUlSS/UlSUl[R "SS[ [R"S55S-SS 9Ul/Ul g) Nbboxsegm)di,ig?gffffff?g!@T)endpoint) ann_fileclassesmetrics proposal_numsnplinspaceintroundiou_thrsresultsselfs u/var/www/html/land-doc-ocr/venv/lib/python3.13/site-packages/modelscope/metrics/image_instance_segmentation_metric.py__init__,ImageInstanceSegmentationCOCOMetric.__init__sY  ' - c"((#456:TK  outputsinputscUSn[US[5(a!UVVs/sHupEU[U54PM nnnURR U5 UR c!USSSUlUSSSUlggs snnf)N eval_resultr img_metasrr) isinstancetupleencode_mask_resultsr"extendrr)r$r)r*result bbox_results mask_resultss r%add'ImageInstanceSegmentationCOCOMetric.add"s' fQi ' '8>@8>4,$%8%FG8> @ F# == #K03J?DM";/29=DL !@sB c[UR5nURURS9UlUR 5UlURURUR 5up#[5nURGHnUnXR;a[US35e[X%5nUS:XaUHnURS5 M URU5n [#XU5n URU R$lUR U R$l[+UR,5U R$lUR0U R$lSSSS S S S S SSSSS. n U R55 U R75 U R95 /SQn U H*n USU 3n[;U R<XS5nXU'M, U R<SS nUSSSUSSSUSSSUS SSUS SSUS S3 XES3'GM UbUR?5 U$![a [!S5  M.f=f)N)catNmsz is not in resultsrrz2The testing results of the whole dataset is empty.rr ) mAPmAP_50mAP_75mAP_smAP_mmAP_lzAR@100zAR@300zAR@1000z AR_s@1000z AR_m@1000z AR_l@1000)rCrDrErFrGrH_z.3f _mAP_copypaste) rr getCatIdsrcat_ids getImgIdsimg_idsformat_resultsr"rrKeyErrorr poploadRes IndexErrorprintrparamscatIdsimgIdslistrmaxDetsr!iouThrsevaluate accumulate summarizefloatstatscleanup)r$cocoGt result_filestmp_dir eval_resultsmetriciou_type predictionsxcocoDtcocoEvalcoco_metric_names metric_items metric_itemkeyvalaps r%r\,ImageInstanceSegmentationCOCOMetric.evaluate-sbdmm$''t||'< '') $ 3 3DLL$,, O "} llFH)&);<== "<#78 v%)f ) 4  9H%)\\HOO "%)\\HOO "&*4+=+=&>HOO #&*mmHOO # !         !    L , +/~~&7&DEcJM$'S! , #Ba5+Qr!uSk2a5+Qr!uSka5+Qr!uSk+ 8>2 3o#t   OO Y JK s!>H,,IIcNURRUR5 gN)r"r1)r$others r%merge)ImageInstanceSegmentationCOCOMetric.mergess EMM*r(cUR$rt)r"r#s r% __getstate__0ImageInstanceSegmentationCOCOMetric.__getstate__vs ||r(c0UR5 Xlgrt)r&r")r$states r% __setstate__0ImageInstanceSegmentationCOCOMetric.__setstate__ys  r(Nc \[U[5(dS5e[U5[U5:Xd)SR[U5[U555eUc7[R "5n[ R"URS5nOSnURX5nXe4$)aFormat the results to json (standard format for COCO evaluation). Args: results (list[tuple | numpy.ndarray]): Testing results of the dataset. data_infos(list[tuple | numpy.ndarray]): data information jsonfile_prefix (str | None): The prefix of json files. It includes the file path and the prefix of filename, e.g., "a/b/prefix". If not specified, a temp file will be created. Default: None. Returns: tuple: (result_files, tmp_dir), result_files is a dict containing the json filepaths, tmp_dir is the temporal directory created for saving json files when jsonfile_prefix is not specified. zresults must be a listz?The length of results is not equal to the dataset len: {} != {}Nr") r.rYlenformattempfileTemporaryDirectoryospjoinname results2json)r$r"rOjsonfile_prefixkwargsrdrcs r%rP2ImageInstanceSegmentationCOCOMetric.format_results}s '4((B*BB(7|s7|+ 0 M F3w<W . 0+  "113G!hhw||Y?OG((B $$r(c^UR5nUSUSUSUS- USUS- /$)zConvert ``xyxy`` style bounding boxes to ``xywh`` style for COCO evaluation. Args: bbox (numpy.ndarray): The bounding boxes, shape (4, ), in ``xyxy`` order. Returns: list[float]: The converted bounding boxes, in ``xywh`` order. rrr9r:)tolist)r$r_bboxs r% xyxy2xywh-ImageInstanceSegmentationCOCOMetric.xyxy2xywhsG  !H !H !HuQx  !HuQx    r(cR/n[[UR55HnURUnXn[URS5HPn[ 5nXGS'UR XV5US'[ XVS5US'SUS'URU5 MR M U$)z,Convert proposal results to COCO json style.rimage_idrr;scorer category_id)rangerrOshapedictrr_append)r$r" json_resultsidximg_idbboxesidatas r%_proposal2json2ImageInstanceSegmentationCOCOMetric._proposal2jsons T\\*+C\\#&F\F6<<?+v#)Z #~~fi8V %fil 3W &']###D) ,,r(c/n[[UR55HnURUnXn[[U55HnU[UR5:a MIXVn[URS5H]n[ 5n XIS'UR Xx5U S'[XxS5U S'URUU S'URU 5 M_ M M U$)z-Convert detection results to COCO json style.rrrr;rr) rrrOrrrrr_rMr) r$r"rrrr2labelrrrs r% _det2json-ImageInstanceSegmentationCOCOMetric._det2jsons T\\*+C\\#&F\Fs6{+ C --v||A/A6D'-$#'>>&)#>&)#>&)#sG)c"[5n[US[5(a1URU5nUS3US'US3US'[ XCS5 U$[US[ 5(aOUR U5nUS3US'US3US'US3US'[ USUS5 [ USUS5 U$[US[R5(a)URU5nUS3US'[ XCS5 U$[S 5e) aDump the detection results to a COCO style json file. There are 3 types of results: proposals, bbox predictions, mask predictions, and they have different data types. This method will automatically recognize the type, and dump them to json files. Args: results (list[list | tuple | ndarray]): Testing results of the dataset. outfile_prefix (str): The filename prefix of the json files. If the prefix is "somepath/xxx", the json files will be named "somepath/xxx.bbox.json", "somepath/xxx.segm.json", "somepath/xxx.proposal.json". Returns: dict[str: str]: Possible keys are "bbox", "segm", "proposal", and values are corresponding filenames. rz .bbox.jsonrproposalz .segm.jsonrrz.proposal.jsonzinvalid type of results) rr.rYrrr/rrndarrayr TypeError)r$r"outfile_prefixrcrs r%r0ImageInstanceSegmentationCOCOMetric.results2jsonsD&v gaj$ ' '>>'2L&4%5Z#@L *8)9'DL $ F3 4 E * *??73L&4%5Z#@L *8)9'DL $&4%5Z#@L a,v"6 7 a,v"6 7  BJJ / /..w7L*8)9'HL $ J7 856 6r()rrMrrOr!rrr")rurrt)__name__ __module__ __qualname____firstlineno____doc__r&rstrrr5r\rvryr}rPrrrrr__static_attributes__r(r%rrsd >4S> >4S> >DL+%: ( .*4X&r(rc [U[5(aUupOUn[U5n[U5Vs/sHn/PM nn[[U55HdnXHYnXVR [ R "[R"USS2SS2[R4SSS95S5 M[ Mf [U[5(aUW4$U$s snf)zEncode bitmap mask to RLE code. Args: mask_results (list | tuple[list]): bitmap mask results. In mask scoring rcnn, mask_results is a tuple of (segm_results, segm_cls_score). Returns: list | tuple: RLE encoded mask. NFuint8)orderdtyper) r.r/rrr mask_utilencoderarraynewaxis)r4 cls_segmscls_mask_scores num_classesrIencoded_mask_resultsrcls_segms r%r0r0)s,&&%1" ? i.K(-k(:;(:1B(:; 3y> "! H # * *  HH Arzz!12#%'()*+ ,%#,&&#_44##rsg#$!)('.3 )J)JLQ&QLQh$r(