j ddlZddlZddlmZddlmZmZ ddlZej sJn##e e e f$rddl mZedddlZYnwxYwddejded efd Zd ed ed ejfdZddededed ejfdZdejded ejfdZdS)N)cdist) batch_probioubbox_ioa)check_requirementsz lap>=0.5.12T cost_matrixthreshuse_lapcjdkrhtjdtt t jdt t jdfS|rrtjd\}dtD}tj dkd}tj dkd}nqtj \tjfdt tD}t|dkrYt!tjjd}t!tjjd}nt!t%tjjdt%|d d dfz }t!t%tjjdt%|d d dfz }|||fS) aPerform linear assignment using either the scipy or lap.lapjv method. Args: cost_matrix (np.ndarray): The matrix containing cost values for assignments, with shape (N, M). thresh (float): Threshold for considering an assignment valid. use_lap (bool): Use lap.lapjv for the assignment. If False, scipy.optimize.linear_sum_assignment is used. Returns: matched_indices (list[list[int]] | np.ndarray): Matched indices of shape (K, 2), where K is the number of matches. unmatched_a (tuple | list | np.ndarray): Unmatched indices from the first set. unmatched_b (tuple | list | np.ndarray): Unmatched indices from the second set. Examples: >>> cost_matrix = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]]) >>> thresh = 5.0 >>> matched_indices, unmatched_a, unmatched_b = linear_assignment(cost_matrix, thresh, use_lap=True) r)rdtypeT) extend_cost cost_limitc(g|]\}}|dk ||gS)r).0ixmxs h/home/longshao/multi-rider-rag/.venv/lib/python3.11/site-packages/ultralytics/trackers/utils/matching.py z%linear_assignment...s%BBBB"''B8'''cdg|],}||fk||g-Srr)rirrxys rrz%linear_assignment..5sHgggq[QRSTQUWXYZW[Q[E\`fEfEfqtQqTlEfEfEfrN)sizenpemptyinttuplerangeshapelaplapjv enumeratewherescipyoptimizelinear_sum_assignmentasarraylenlistarange frozenset) rrr _matches unmatched_a unmatched_brrs `` @@rlinear_assignmentr4s&1xc***E% 8I!8L2M2M,N,NPUV[\g\mno\pVqVqPrPrrrf)KTfMMM1aBB)A,,BBBhq1uooa( hq1uooa( ~33K@@1*gggggggE#a&&MMggghh w<<1  ry):1)=>>??Kry):1)=>>??KKy;3DQ3G)H)HIIIV]^_^_^_ab^bVcLdLddeeKy;3DQ3G)H)HIIIV]^_^_^_ab^bVcLdLddeeK K ,,ratracksbtracksreturnc@|r t|dtjs"|r%t|dtjr|}|}nd|D}d|D}tjt |t |ftj}t |rt |rt |ddkrxt |ddkr_t tj|tjtj|tj}nNttj|tjtj|tjd}d|z S) aCompute cost based on Intersection over Union (IoU) between tracks. Args: atracks (list[STrack] | list[np.ndarray]): List of tracks 'a' or bounding boxes. btracks (list[STrack] | list[np.ndarray]): List of tracks 'b' or bounding boxes. Returns: (np.ndarray): Cost matrix computed based on IoU with shape (len(atracks), len(btracks)). Examples: Compute IoU distance between two sets of tracks >>> atracks = [np.array([0, 0, 10, 10]), np.array([20, 20, 30, 30])] >>> btracks = [np.array([5, 5, 15, 15]), np.array([25, 25, 35, 35])] >>> cost_matrix = iou_distance(atracks, btracks) rc8g|]}|j|jn|jSNanglexywhaxyxyrtracks rrz iou_distance..T'^^^U!8%++ej^^^rc8g|]}|j|jn|jSr:r;r?s rrz iou_distance..UrArr T)iour) isinstancerndarrayzerosr,float32rascontiguousarraynumpyr)r5r6atlbrsbtlbrsiouss r iou_distancerN@s _Jwqz2:66_G_ SZ[\S]_a_iHjHj_^^V]^^^^^V]^^^ 8S[[#f++.bj A A AD 6{{ s6{{  vay>>Q  3vay>>Q#6#6 $V2:>>>$V2:>>>egg D $V2:>>>$V2:>>>D t8Orcosinetracks detectionsmetricctjt|t|ftj}|jdkr|Stjd|Dtj}tjd|Dtj}tjdt|||}|S)aCompute distance between tracks and detections based on embeddings. Args: tracks (list[BOTrack]): List of tracks, where each track contains embedding features. detections (list[BOTrack]): List of detections, where each detection contains embedding features. metric (str): Metric for distance computation. Supported metrics include 'cosine', 'euclidean', etc. Returns: (np.ndarray): Cost matrix computed based on embeddings with shape (N, M), where N is the number of tracks and M is the number of detections. Examples: Compute the embedding distance between tracks and detections using cosine metric >>> tracks = [BOTrack(...), BOTrack(...)] # List of track objects with embedding features >>> detections = [BOTrack(...), BOTrack(...)] # List of detection objects with embedding features >>> cost_matrix = embedding_distance(tracks, detections, metric="cosine") r rcg|] }|j Sr) curr_featr?s rrz&embedding_distance..|sGGG5uGGGrcg|] }|j Sr) smooth_featr?s rrz&embedding_distance..s G G Gu!2 G G Grg)rrGr,rHrr+maximumr)rPrQrRr det_featurestrack_featuress rembedding_distancer[gs$(CKKZ9LLLK1:GGJGGGrzZZZLZ G G G G GrzZZZN*S% f"M"MNNK rc|jdkr|Sd|z }tjd|D}|d|jdd}||z}d|z S)aFuse cost matrix with detection scores to produce a single cost matrix. Args: cost_matrix (np.ndarray): The matrix containing cost values for assignments, with shape (N, M). detections (list[BaseTrack]): List of detections, each containing a score attribute. Returns: (np.ndarray): Fused cost matrix with shape (N, M). Examples: Fuse a cost matrix with detection scores >>> cost_matrix = np.random.rand(5, 10) # 5 tracks and 10 detections >>> detections = [BaseTrack(score=np.random.rand()) for _ in range(10)] >>> fused_matrix = fuse_score(cost_matrix, detections) rrcg|] }|j Sr)score)rdets rrzfuse_score..s;;;39;;;rN)axis)rrarrayrepeatr#)rrQiou_sim det_scoresfuse_sims r fuse_scorerfsx 1+oG;; ;;;<rqs ((((((========JJJ ??^^4;;;;;;}%%%JJJJJ )-)-2:)-u)-t)-)-)-)-X$$$$"*$$$$NtsRTR\:BJDRZs &AA