القائمة
Encode motion and velocity by computing per-channel pixel differences between consecutive frames instead of stacking raw frames for RL visual observations
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
Memory-efficient Keras generator that streams sharded CSV files in chunks, renders strokes to images on-the-fly, and yields batches for training on datasets too large for memory
Stack 1D convolutions for local feature extraction before bidirectional LSTMs to classify variable-length stroke sequences into hundreds of doodle categories
Partition a large dataset into N balanced shards using integer key modulo arithmetic for reproducible, class-interleaved splits across CSV files
Normalize raw stroke coordinates to 0-255 range, resample at uniform arc-length spacing, then apply Ramer-Douglas-Peucker simplification
Render stroke sequences to grayscale images with temporal intensity encoding where earlier strokes are brighter and later strokes fade to encode drawing order
Score each face in a video as anomalous by computing L2 distance from the embedding centroid of all faces, then convert to probability via logistic function
| name | cv-frame-differencing-temporal-encoding |
| description | Encode motion and velocity by computing per-channel pixel differences between consecutive frames instead of stacking raw frames for RL visual observations |
Frame stacking (concatenating the last N frames) is the standard way to give RL agents temporal information from visual observations. Frame differencing is a lighter alternative: subtract the previous frame from the current one to produce a motion-only image. Moving objects appear as non-zero pixels while static backgrounds cancel out, giving the network explicit velocity signals without doubling the input channels.
import numpy as np
from collections import deque
class FrameDiffWrapper:
def __init__(self, env, n_channels=4):
self.env = env
self.n_channels = n_channels
self.buffer = deque(maxlen=2)
def reset(self):
obs = self.env.reset()
frame = obs / 255.0
self.buffer.append(frame)
self.buffer.append(np.zeros_like(frame))
return self._diff()
def step(self, action):
obs, reward, done, info = self.env.step(action)
self.buffer.append(obs / 255.0)
return self._diff(), reward, done, info
def _diff(self):
diff = np.empty_like(self.buffer[1])
for c in range(diff.shape[-1]):
diff[..., c] = self.buffer[1][..., c] - self.buffer[0][..., c]
return diff
current - previous