miner.py

from pathlib import Path
import math

import cv2
import numpy as np
import onnxruntime as ort
from numpy import ndarray
from pydantic import BaseModel


class BoundingBox(BaseModel):
    x1: int
    y1: int
    x2: int
    y2: int
    cls_id: int
    conf: float


class TVFrameResult(BaseModel):
    frame_id: int
    boxes: list[BoundingBox]
    keypoints: list[tuple[int, int]]


class Miner:
    """ONNX Runtime miner for beverage detection (cup / bottle / can).

    Strategy (ported from fire001 miner):
      - per-class confidence threshold with per-class rescue bonus
      - per-class hard NMS, then cross-class dedup (margin-ordered)
      - horizontal-flip TTA with class-aware cluster-max score boost
      - sanity-box filter for tiny / spanning / extreme-AR boxes
    """

    # Output / canonical class order. This is what every downstream consumer
    # (validator, evaluator, BoundingBox.cls_id) sees.
    class_names = ["cup", "bottle", "can"]
    # FALLBACK order the model emits classes in -- remapped to `class_names`
    # index by `self.cls_remap` (built in __init__). The authoritative order
    # is read from the ONNX `names` metadata that Ultralytics embeds at
    # export time (self-contained in weights.onnx, no external data), so a
    # retrained model with a different class order is remapped correctly
    # without code changes. This list is used only when that metadata is
    # missing or unparsable. Beverage-13-trained models emit bottle/can/cup.
    _model_class_order = ["bottle", "can", "cup"]

    # FALLBACK input size, used only when the ONNX input shape is dynamic.
    # The actual size is read from the model (fixed-shape exports), so this
    # has no effect on the current weights.onnx (fixed 1280) and the new
    # validator-matched models (fixed 1024) alike.
    input_size = 1024
    # NMS IoU. 0.45 measured best on the validator-style val split (282
    # 1024x1024 crops, composite = 0.6*mAP50 + 0.4*FP-pillar) via
    # tune_miner.py -- re-run that sweep after any model retrain.
    iou_thres = 0.45
    cross_iou_thresh = 0.8
    # Containment (intersection-over-minimum-area, aka IoMin) threshold.
    # Plain IoU misses the "big box + small box on the SAME object" case: a
    # small box fully nested inside a large one has IoU ~= small/large (often
    # < 0.5) yet ~1.0 containment, so IoU-only NMS leaves both boxes. When the
    # smaller box's overlap with the larger exceeds this fraction, the pair is
    # treated as one object and the lower-scoring box is suppressed. Lower it
    # toward ~0.7 if duplicates still leak; raise it toward 0.9 if it merges
    # genuinely distinct, partially-occluding drinks.
    containment_thresh = 0.8
    min_side = 12.0
    min_box_area = 100.0
    max_aspect_ratio = 10.0
    max_det = 150

    # Per-class confidence thresholds. Indexed by canonical class_names order:
    # [cup, bottle, can]. Cup is highest because paper / plastic cups blur
    # into hands and skin in low-bitrate CCTV. Values are the measured
    # optimum of a full grid sweep on the validator-style val split
    # (tune_miner.py, composite 0.7104 -> 0.7143 over the previous
    # [0.58, 0.45, 0.42]) -- re-run the sweep after any model retrain.
    _conf_thres_array = np.array([0.55, 0.50, 0.45], dtype=np.float32)
    # Per-class rescue bonus, also indexed by canonical class_names order.
    # If a class has ZERO boxes passing its threshold in a frame, its top-1
    # candidate is admitted when its score is at least (threshold - bonus).
    # DISABLED (all zeros): the same sweep showed rescue admits more false
    # positives than true positives under the validator's FP pillar
    # (0.4 weight); each rescued borderline box costs more than it gains.
    _bonus_array = np.array([0.0, 0.0, 0.0], dtype=np.float32)

    def __init__(self, path_hf_repo: Path) -> None:
        model_path = path_hf_repo / "weights.onnx"
        print("ORT version:", ort.__version__)

        try:
            ort.preload_dlls()
            print("preload_dlls success")
        except Exception as e:
            print(f"preload_dlls failed: {e}")

        print("ORT available providers BEFORE session:", ort.get_available_providers())

        sess_options = ort.SessionOptions()
        sess_options.graph_optimization_level = ort.GraphOptimizationLevel.ORT_ENABLE_ALL

        try:
            self.session = ort.InferenceSession(
                str(model_path),
                sess_options=sess_options,
                providers=["CUDAExecutionProvider", "CPUExecutionProvider"],
            )
            print("Created ORT session with preferred CUDA provider list")
        except Exception as e:
            print(f"CUDA session creation failed, falling back to CPU: {e}")
            self.session = ort.InferenceSession(
                str(model_path),
                sess_options=sess_options,
                providers=["CPUExecutionProvider"],
            )

        print("ORT session providers:", self.session.get_providers())

        # Build cls_remap: for each model-emit index i,
        #   cls_remap[i] = self.class_names.index(model_class_order[i])
        # i.e. converts a model-side class id into the canonical class id
        # that downstream code (BoundingBox.cls_id, validator) expects.
        # The model-side order comes from the ONNX metadata when available
        # (authoritative -- embedded by Ultralytics at export, ships inside
        # weights.onnx), else falls back to the static _model_class_order.
        model_class_order = self._read_model_class_order()
        if model_class_order is None:
            model_class_order = list(self._model_class_order)
            print(f"cls order: no usable ONNX metadata, FALLBACK {model_class_order}")
        else:
            print(f"cls order: from ONNX metadata {model_class_order}")
        self.cls_remap = np.array(
            [self.class_names.index(n) for n in model_class_order],
            dtype=np.int32,
        )

        for inp in self.session.get_inputs():
            print("INPUT:", inp.name, inp.shape, inp.type)
        for out in self.session.get_outputs():
            print("OUTPUT:", out.name, out.shape, out.type)

        self.input_name = self.session.get_inputs()[0].name
        self.output_names = [output.name for output in self.session.get_outputs()]
        self.input_shape = self.session.get_inputs()[0].shape

        self.input_height = self._safe_dim(self.input_shape[2], default=self.input_size)
        self.input_width = self._safe_dim(self.input_shape[3], default=self.input_size)

        self.use_tta = True

        print(f"ONNX model loaded from: {model_path}")
        print(f"ONNX input: name={self.input_name}, shape={self.input_shape}")
        print("per-class conf: " + ", ".join(
            f"{n}={t:.3f}" for n, t in zip(self.class_names,
                                           self._conf_thres_array.tolist())))

    def __repr__(self) -> str:
        return (
            f"ONNXRuntime(session={type(self.session).__name__}, "
            f"providers={self.session.get_providers()})"
        )

    @staticmethod
    def _safe_dim(value, default: int) -> int:
        return value if isinstance(value, int) and value > 0 else default

    def _read_model_class_order(self) -> list[str] | None:
        """Read the model's class order from Ultralytics ONNX metadata.

        Returns the class names ordered by model-emit index, or None when
        metadata is missing/unparsable or doesn't match `class_names` as a
        set (in which case the static _model_class_order fallback is used).
        """
        try:
            import ast

            meta = self.session.get_modelmeta().custom_metadata_map
            names = ast.literal_eval(meta["names"])  # e.g. {0: 'cup', 1: ...}
            if isinstance(names, dict):
                order = [str(names[i]) for i in sorted(names)]
            else:
                order = [str(n) for n in names]
        except Exception as e:
            print(f"cls order: could not read ONNX names metadata ({e})")
            return None
        if sorted(order) != sorted(self.class_names):
            print(
                f"cls order: ONNX names {order} do not match expected classes "
                f"{self.class_names}; ignoring metadata"
            )
            return None
        return order

    def _letterbox(self, image: ndarray, new_shape: tuple[int, int],
                   color=(114, 114, 114)
                   ) -> tuple[ndarray, float, tuple[float, float]]:
        h, w = image.shape[:2]
        new_w, new_h = new_shape
        ratio = min(new_w / w, new_h / h)
        resized_w = int(round(w * ratio))
        resized_h = int(round(h * ratio))
        if (resized_w, resized_h) != (w, h):
            interp = cv2.INTER_CUBIC if ratio > 1.0 else cv2.INTER_LINEAR
            image = cv2.resize(image, (resized_w, resized_h), interpolation=interp)
        dw = (new_w - resized_w) / 2.0
        dh = (new_h - resized_h) / 2.0
        left = int(round(dw - 0.1))
        right = int(round(dw + 0.1))
        top = int(round(dh - 0.1))
        bottom = int(round(dh + 0.1))
        padded = cv2.copyMakeBorder(image, top, bottom, left, right,
                                    borderType=cv2.BORDER_CONSTANT, value=color)
        return padded, ratio, (dw, dh)

    def _preprocess(self, image: ndarray
                    ) -> tuple[np.ndarray, float, tuple[float, float],
                               tuple[int, int]]:
        orig_h, orig_w = image.shape[:2]
        img, ratio, pad = self._letterbox(image, (self.input_width, self.input_height))
        img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
        img = img.astype(np.float32) / 255.0
        img = np.transpose(img, (2, 0, 1))[None, ...]
        img = np.ascontiguousarray(img, dtype=np.float32)
        return img, ratio, pad, (orig_w, orig_h)

    @staticmethod
    def _clip_boxes(boxes: np.ndarray, image_size: tuple[int, int]) -> np.ndarray:
        w, h = image_size
        boxes[:, 0] = np.clip(boxes[:, 0], 0, w - 1)
        boxes[:, 1] = np.clip(boxes[:, 1], 0, h - 1)
        boxes[:, 2] = np.clip(boxes[:, 2], 0, w - 1)
        boxes[:, 3] = np.clip(boxes[:, 3], 0, h - 1)
        return boxes

    @staticmethod
    def _xywh_to_xyxy(boxes: np.ndarray) -> np.ndarray:
        out = np.empty_like(boxes)
        out[:, 0] = boxes[:, 0] - boxes[:, 2] / 2.0
        out[:, 1] = boxes[:, 1] - boxes[:, 3] / 2.0
        out[:, 2] = boxes[:, 0] + boxes[:, 2] / 2.0
        out[:, 3] = boxes[:, 1] + boxes[:, 3] / 2.0
        return out

    def _hard_nms(self, boxes: np.ndarray, scores: np.ndarray,
                  iou_thresh: float) -> np.ndarray:
        n = len(boxes)
        if n == 0:
            return np.array([], dtype=np.intp)
        order = np.argsort(-scores)
        keep: list[int] = []
        while len(order) > 0:
            i = int(order[0])
            keep.append(i)
            if len(order) == 1:
                break
            rest = order[1:]
            xx1 = np.maximum(boxes[i, 0], boxes[rest, 0])
            yy1 = np.maximum(boxes[i, 1], boxes[rest, 1])
            xx2 = np.minimum(boxes[i, 2], boxes[rest, 2])
            yy2 = np.minimum(boxes[i, 3], boxes[rest, 3])
            inter = np.maximum(0.0, xx2 - xx1) * np.maximum(0.0, yy2 - yy1)
            a_i = (max(0.0, boxes[i, 2] - boxes[i, 0]) *
                   max(0.0, boxes[i, 3] - boxes[i, 1]))
            a_r = (np.maximum(0.0, boxes[rest, 2] - boxes[rest, 0]) *
                   np.maximum(0.0, boxes[rest, 3] - boxes[rest, 1]))
            iou = inter / (a_i + a_r - inter + 1e-7)
            # Containment (IoMin): catches a small box nested in a large one,
            # where IoU is low but the smaller box is almost entirely covered.
            containment = inter / (np.minimum(a_i, a_r) + 1e-7)
            suppress = (iou > iou_thresh) | (containment > self.containment_thresh)
            order = rest[~suppress]
        return np.array(keep, dtype=np.intp)

    def _per_class_hard_nms(self, boxes: np.ndarray, scores: np.ndarray,
                            cls_ids: np.ndarray, iou_thresh: float
                            ) -> np.ndarray:
        if len(boxes) == 0:
            return np.array([], dtype=np.intp)
        all_keep: list[int] = []
        for c in np.unique(cls_ids):
            mask = cls_ids == c
            indices = np.where(mask)[0]
            keep = self._hard_nms(boxes[mask], scores[mask], iou_thresh)
            all_keep.extend(indices[keep].tolist())
        all_keep.sort()
        return np.array(all_keep, dtype=np.intp)

    def _conf_filter_mask(self, scores: np.ndarray,
                          cls_ids: np.ndarray) -> np.ndarray:
        """Boolean keep-mask: score >= per-class threshold, with a per-class
        rescue -- if a class has zero boxes passing, admit its top-1 candidate
        when its score >= (per-class threshold - per-class bonus).

        This recovers the common failure mode where one class is genuinely
        present in the frame but every candidate sits just below its
        threshold (e.g. a single faint can in a stadium concourse shot).
        """
        if len(scores) == 0:
            return np.zeros(0, dtype=bool)
        thr = self._conf_thres_array[cls_ids]
        keep = scores >= thr
        for c in np.unique(cls_ids):
            b = float(self._bonus_array[c])
            if b <= 0.0:
                continue
            cm = cls_ids == c
            if keep[cm].any():
                continue
            idx = np.where(cm)[0]
            top = int(idx[int(np.argmax(scores[idx]))])
            if scores[top] >= self._conf_thres_array[c] - b:
                keep[top] = True
        return keep

    def _cross_class_dedup_op(self, boxes: np.ndarray, scores: np.ndarray,
                              cls_ids: np.ndarray, iou_thresh: float
                              ) -> tuple[np.ndarray, np.ndarray, np.ndarray]:
        """Remove near-duplicate boxes across classes.

        Order candidates by (score - per_class_threshold) margin, then by
        area; keep the highest-margin, suppress every other box with IoU
        > iou_thresh. Margin ordering matters here because cup / bottle /
        can use different thresholds: a bottle at 0.45 (margin +0.15 over
        its 0.30 threshold) is more confident than a cup at 0.55 (margin
        +0.05 over its 0.50 threshold), even though the raw score is lower.
        """
        n = len(boxes)
        if n <= 1:
            return boxes, scores, cls_ids
        boxes = np.asarray(boxes, dtype=np.float32)
        scores = np.asarray(scores, dtype=np.float32)
        cls_ids = np.asarray(cls_ids, dtype=np.int32)
        areas = (np.maximum(0.0, boxes[:, 2] - boxes[:, 0]) *
                 np.maximum(0.0, boxes[:, 3] - boxes[:, 1]))
        margins = scores - self._conf_thres_array[cls_ids]
        order = np.lexsort((-areas, -margins))
        suppressed = np.zeros(n, dtype=bool)
        keep: list[int] = []
        for i in order:
            if suppressed[i]:
                continue
            keep.append(int(i))
            bi = boxes[i]
            xx1 = np.maximum(bi[0], boxes[:, 0])
            yy1 = np.maximum(bi[1], boxes[:, 1])
            xx2 = np.minimum(bi[2], boxes[:, 2])
            yy2 = np.minimum(bi[3], boxes[:, 3])
            inter = np.maximum(0.0, xx2 - xx1) * np.maximum(0.0, yy2 - yy1)
            a_i = max(1e-7, float((bi[2] - bi[0]) * (bi[3] - bi[1])))
            iou = inter / (a_i + areas - inter + 1e-7)
            # Containment (IoMin): suppress a smaller box mostly covered by the
            # kept one even when IoU is low (large box + nested small box on
            # the same physical object, possibly a different class).
            containment = inter / (np.minimum(a_i, areas) + 1e-7)
            dup = (iou > iou_thresh) | (containment > self.containment_thresh)
            dup[i] = False
            suppressed |= dup
        keep_idx = np.array(keep, dtype=np.intp)
        return boxes[keep_idx], scores[keep_idx], cls_ids[keep_idx]

    def _filter_sane_boxes(self, boxes: np.ndarray, scores: np.ndarray,
                           cls_ids: np.ndarray, orig_size: tuple[int, int]
                           ) -> tuple[np.ndarray, np.ndarray, np.ndarray]:
        if len(boxes) == 0:
            return boxes, scores, cls_ids
        orig_w, orig_h = orig_size
        image_area = float(orig_w * orig_h)
        bw = np.maximum(0.0, boxes[:, 2] - boxes[:, 0])
        bh = np.maximum(0.0, boxes[:, 3] - boxes[:, 1])
        area = bw * bh
        ar = np.where(
            (bw > 0) & (bh > 0),
            np.maximum(bw / np.maximum(bh, 1e-6), bh / np.maximum(bw, 1e-6)),
            np.inf,
        )
        keep = (
            (bw >= self.min_side) & (bh >= self.min_side) &
            (area >= self.min_box_area) &
            (area <= 0.95 * image_area) &
            (ar <= self.max_aspect_ratio)
        )
        return boxes[keep], scores[keep], cls_ids[keep]

    @staticmethod
    def _max_score_per_cluster(post_boxes: np.ndarray,
                               post_cls: np.ndarray,
                               full_boxes: np.ndarray,
                               full_scores: np.ndarray,
                               full_cls: np.ndarray,
                               iou_thresh: float) -> np.ndarray:
        """For each kept (post-NMS) box, return the max score over the FULL
        candidate set among SAME-CLASS boxes with IoU >= iou_thresh.

        The previous version omitted the same-class constraint, which let a
        confident bottle raise the score of a coincident cup (or vice versa)
        under TTA -- a silent FP booster. Fixed here.
        """
        n = len(post_boxes)
        if n == 0:
            return np.empty(0, dtype=np.float32)
        full_areas = (np.maximum(0.0, full_boxes[:, 2] - full_boxes[:, 0]) *
                      np.maximum(0.0, full_boxes[:, 3] - full_boxes[:, 1]))
        out = np.empty(n, dtype=np.float32)
        for i in range(n):
            bi = post_boxes[i]
            xx1 = np.maximum(bi[0], full_boxes[:, 0])
            yy1 = np.maximum(bi[1], full_boxes[:, 1])
            xx2 = np.minimum(bi[2], full_boxes[:, 2])
            yy2 = np.minimum(bi[3], full_boxes[:, 3])
            inter = np.maximum(0.0, xx2 - xx1) * np.maximum(0.0, yy2 - yy1)
            a_i = max(0.0, float((bi[2] - bi[0]) * (bi[3] - bi[1])))
            iou = inter / (a_i + full_areas - inter + 1e-7)
            cluster = (iou >= iou_thresh) & (full_cls == post_cls[i])
            out[i] = float(np.max(full_scores[cluster])) if np.any(cluster) else 0.0
        return out

    def _per_view_pipeline(self, boxes: np.ndarray, scores: np.ndarray,
                           cls_ids: np.ndarray, orig_size: tuple[int, int]
                           ) -> tuple[np.ndarray, np.ndarray, np.ndarray]:
        """Per-view post-processing: sanity filter -> per-class NMS -> cap
        -> cross-class dedup.

        Uses per-class NMS so an overlapping cup-and-can pair survives the
        first dedup pass; the cross-class dedup at the tail then resolves
        any genuinely-same-object collision using the (score - threshold)
        margin ordering.
        """
        boxes, scores, cls_ids = self._filter_sane_boxes(
            boxes, scores, cls_ids, orig_size
        )
        if len(boxes) == 0:
            return boxes, scores, cls_ids
        if len(boxes) > 1:
            keep = self._per_class_hard_nms(boxes, scores, cls_ids, self.iou_thres)
            boxes, scores, cls_ids = boxes[keep], scores[keep], cls_ids[keep]
        if len(scores) > self.max_det:
            top = np.argsort(-scores)[: self.max_det]
            boxes, scores, cls_ids = boxes[top], scores[top], cls_ids[top]
        if len(boxes) > 1:
            boxes, scores, cls_ids = self._cross_class_dedup_op(
                boxes, scores, cls_ids, self.cross_iou_thresh
            )
        return boxes, scores, cls_ids

    def _decode_final_dets(self, preds: np.ndarray, ratio: float,
                           pad: tuple[float, float],
                           orig_size: tuple[int, int]) -> list[BoundingBox]:
        if preds.ndim == 3 and preds.shape[0] == 1:
            preds = preds[0]
        if preds.ndim != 2 or preds.shape[1] < 6:
            raise ValueError(f"Unexpected ONNX final-det output shape: {preds.shape}")

        boxes = preds[:, :4].astype(np.float32)
        scores = preds[:, 4].astype(np.float32)
        cls_ids = preds[:, 5].astype(np.int32)
        # Remap model-emit indices to canonical class_names indices.
        cls_ids = self.cls_remap[cls_ids]

        keep = self._conf_filter_mask(scores, cls_ids)
        boxes = boxes[keep]
        scores = scores[keep]
        cls_ids = cls_ids[keep]
        if len(boxes) == 0:
            return []

        pad_w, pad_h = pad
        boxes[:, [0, 2]] -= pad_w
        boxes[:, [1, 3]] -= pad_h
        boxes /= ratio
        boxes = self._clip_boxes(boxes, orig_size)

        boxes, scores, cls_ids = self._per_view_pipeline(
            boxes, scores, cls_ids, orig_size
        )
        return self._build_results(boxes, scores, cls_ids)

    def _decode_raw_yolo(self, preds: np.ndarray, ratio: float,
                         pad: tuple[float, float],
                         orig_size: tuple[int, int]) -> list[BoundingBox]:
        if preds.ndim != 3 or preds.shape[0] != 1:
            raise ValueError(f"Unexpected raw ONNX output shape: {preds.shape}")
        preds = preds[0]
        if preds.shape[0] <= 16 and preds.shape[1] > preds.shape[0]:
            preds = preds.T
        if preds.ndim != 2 or preds.shape[1] < 5:
            raise ValueError(f"Unexpected raw output shape: {preds.shape}")

        boxes_xywh = preds[:, :4].astype(np.float32)
        cls_part = preds[:, 4:].astype(np.float32)
        if cls_part.shape[1] == 1:
            scores = cls_part[:, 0]
            cls_ids = np.zeros(len(scores), dtype=np.int32)
        else:
            cls_ids = np.argmax(cls_part, axis=1).astype(np.int32)
            scores = cls_part[np.arange(len(cls_part)), cls_ids]
        # Remap model-emit indices to canonical class_names indices.
        cls_ids = self.cls_remap[cls_ids]

        keep = self._conf_filter_mask(scores, cls_ids)
        boxes_xywh = boxes_xywh[keep]
        scores = scores[keep]
        cls_ids = cls_ids[keep]
        if len(boxes_xywh) == 0:
            return []
        boxes = self._xywh_to_xyxy(boxes_xywh)

        pad_w, pad_h = pad
        boxes[:, [0, 2]] -= pad_w
        boxes[:, [1, 3]] -= pad_h
        boxes /= ratio
        boxes = self._clip_boxes(boxes, orig_size)

        boxes, scores, cls_ids = self._per_view_pipeline(
            boxes, scores, cls_ids, orig_size
        )
        return self._build_results(boxes, scores, cls_ids)

    @staticmethod
    def _build_results(boxes: np.ndarray, scores: np.ndarray,
                       cls_ids: np.ndarray) -> list[BoundingBox]:
        results: list[BoundingBox] = []
        for box, conf, cls_id in zip(boxes, scores, cls_ids):
            x1, y1, x2, y2 = box.tolist()
            if x2 <= x1 or y2 <= y1:
                continue
            results.append(
                BoundingBox(
                    x1=int(math.floor(x1)),
                    y1=int(math.floor(y1)),
                    x2=int(math.ceil(x2)),
                    y2=int(math.ceil(y2)),
                    cls_id=int(cls_id),
                    conf=float(conf),
                )
            )
        return results

    def _postprocess(self, output: np.ndarray, ratio: float,
                     pad: tuple[float, float],
                     orig_size: tuple[int, int]) -> list[BoundingBox]:
        if output.ndim == 2 and output.shape[1] >= 6:
            return self._decode_final_dets(output, ratio, pad, orig_size)
        if output.ndim == 3 and output.shape[0] == 1 and output.shape[2] == 6:
            return self._decode_final_dets(output, ratio, pad, orig_size)
        return self._decode_raw_yolo(output, ratio, pad, orig_size)

    def _predict_single(self, image: np.ndarray) -> list[BoundingBox]:
        if image is None:
            raise ValueError("Input image is None")
        if not isinstance(image, np.ndarray):
            raise TypeError(f"Input is not numpy array: {type(image)}")
        if image.ndim != 3:
            raise ValueError(f"Expected HWC image, got shape={image.shape}")
        if image.shape[2] != 3:
            raise ValueError(f"Expected 3 channels, got shape={image.shape}")
        if image.dtype != np.uint8:
            image = image.astype(np.uint8)

        input_tensor, ratio, pad, orig_size = self._preprocess(image)
        expected = (1, 3, self.input_height, self.input_width)
        if input_tensor.shape != expected:
            raise ValueError(
                f"Bad input tensor shape={input_tensor.shape}, expected={expected}"
            )

        outputs = self.session.run(self.output_names, {self.input_name: input_tensor})
        return self._postprocess(outputs[0], ratio, pad, orig_size)

    def _predict_tta(self, image: np.ndarray) -> list[BoundingBox]:
        """Horizontal-flip TTA.

        Strategy (ported from fire001):
          1. Predict on original and on flipped image.
          2. Map flipped boxes back to original coordinates.
          3. Per-class hard NMS on the union.
          4. For each kept box, compute the max SAME-CLASS score across the
             FULL union -- a high-confidence flipped detection raises a
             borderline original one, but never one of a different class.
          5. Cross-class dedup (margin-ordered) on the kept set to suppress
             same-physical-object multi-class collisions.
        """
        boxes_orig = self._predict_single(image)
        flipped = cv2.flip(image, 1)
        boxes_flip = self._predict_single(flipped)
        w = image.shape[1]
        boxes_flip = [
            BoundingBox(
                x1=w - b.x2, y1=b.y1, x2=w - b.x1, y2=b.y2,
                cls_id=b.cls_id, conf=b.conf,
            )
            for b in boxes_flip
        ]
        all_boxes = boxes_orig + boxes_flip
        if not all_boxes:
            return []

        coords = np.array(
            [[b.x1, b.y1, b.x2, b.y2] for b in all_boxes], dtype=np.float32
        )
        scores = np.array([b.conf for b in all_boxes], dtype=np.float32)
        cls_ids = np.array([b.cls_id for b in all_boxes], dtype=np.int32)

        hard_keep = self._per_class_hard_nms(coords, scores, cls_ids, self.iou_thres)
        if len(hard_keep) == 0:
            return []
        if len(hard_keep) > self.max_det:
            top = np.argsort(-scores[hard_keep])[: self.max_det]
            hard_keep = hard_keep[top]

        # Class-aware cluster-max score boost (fixes the silent cross-class
        # leak in the previous _max_score_per_cluster).
        boosted = self._max_score_per_cluster(
            coords[hard_keep], cls_ids[hard_keep],
            coords, scores, cls_ids, self.iou_thres,
        )

        kept_coords = coords[hard_keep]
        kept_cls = cls_ids[hard_keep]
        if len(kept_coords) > 1:
            kept_coords, boosted, kept_cls = self._cross_class_dedup_op(
                kept_coords, boosted, kept_cls, self.cross_iou_thresh
            )

        return [
            BoundingBox(
                x1=int(math.floor(kept_coords[j, 0])),
                y1=int(math.floor(kept_coords[j, 1])),
                x2=int(math.ceil(kept_coords[j, 2])),
                y2=int(math.ceil(kept_coords[j, 3])),
                cls_id=int(kept_cls[j]),
                conf=float(boosted[j]),
            )
            for j in range(len(kept_coords))
        ]

    def predict_batch(self, batch_images: list[ndarray], offset: int,
                      n_keypoints: int) -> list[TVFrameResult]:
        results: list[TVFrameResult] = []
        for frame_number_in_batch, image in enumerate(batch_images):
            try:
                if self.use_tta:
                    boxes = self._predict_tta(image)
                else:
                    boxes = self._predict_single(image)
            except Exception as e:
                print(f"Inference failed for frame {offset + frame_number_in_batch}: {e}")
                boxes = []
            results.append(
                TVFrameResult(
                    frame_id=offset + frame_number_in_batch,
                    boxes=boxes,
                    keypoints=[(0, 0) for _ in range(max(0, int(n_keypoints)))],
                )
            )
        return results