ZERO-SHOT OBJECT DETECTION AND SEGMENTATION FOR CONSTRUCTION SITES THROUGH MULTI-MODEL INTEGRATION
Bibliographic record
Abstract
Object detection and segmentation are crucial for managing construction sites, aiding in tasks such as progress tracking, material management, and safety assurance.However, conventional methods encounter persistent challenges, including occlusion, variable lighting conditions, and the labor-intensive nature of dataset creation, which limit their adaptability to dynamic construction environments.This study introduces a novel zero-shot object detection and segmentation framework designed specifically for construction-related objects, including machinery, workers, and materials.The proposed framework integrates three state-of-the-art models: Florence-2, Llama3.2-Vision, and the Segment Anything Model 2 (SAM2).Florence-2 generates region proposals for previously unseen objects using textual descriptions; Llama3.2-Visionpredicts and refines accurate labels for detected regions based on textual queries; and SAM2 produces high-precision segmentation masks.The effectiveness of this approach was validated through both qualitative and quantitative experiments.While parts of this framework and qualitative experiments were previously presented, this paper extends our previous work by providing a more detailed methodology and including additional quantitative experiments.Qualitative experiments using images from a specific tunnel excavation site, demonstrating robust detection and segmentation performance under challenging conditions such as occlusion and variable lighting.Quantitative experiments using the Alberta Construction Image Dataset (ACID) showed that the proposed multimodel method significantly outperformed Florence-2 alone, particularly for large objects, despite not achieving the accuracy of the fine-tuned YOLOv11 model.The proposed framework eliminates the need for extensive retraining and manual dataset creation by leveraging the complementary strengths of these models.This scalable and flexible solution offers practical applications in progress tracking, material management, and safety monitoring and thereby addresses the unique complexities of dynamic construction environments.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.003 |
| Bibliometrics | 0.003 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.004 | 0.003 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".