Scene Classification in Indoor Environments for Robots using Context Based Word Embeddings

Bao Xin Chen; Dekun Wu; John K. Tsotsos; Manos Papagelis; Raghavender Sahdev; Xing Zhao

REVIEW 3 major objections 6 minor 26 references

Scene Classification in Indoor Environments for Robots using Context Based Word Embeddings

T0 review · 3 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper claims that re-ranking a CNN's top-5 scene predictions by the cosine similarity between detected objects and scene labels, under a GPS-based environment taxonomy, improves indoor scene-classification accuracy for robots on…

desk verdict Plausible word2vec re-ranking for indoor scene classification, but the reported 1-2 point gains are not statistically supported by the autocorrelated test set. read the letter →

arxiv 1908.06422 v1 pith:GFS3FKAR submitted 2019-08-18 cs.RO cs.CV

Bao Xin Chen , Raghavender Sahdev , Dekun Wu , Xing Zhao , Manos Papagelis , John K. Tsotsos This is my paper · ORCID

classification cs.ROcs.CV

keywords sceneclassificationindoorroboticswordembeddingsword2vecconvolutionalneuralnetworkcosinesimilarityPlaces365environmenttaxonomy

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to make indoor scene classification accurate enough for a robot to use, by combining two signals: a CNN's top-5 scene guesses and the objects visible in the image. The central claim is that comparing word-embedding vectors of the detected objects with vectors of each candidate scene label can re-rank those top-5 guesses and push accuracy above the CNN alone. On the authors' own real-world test set, the refinement adds about 1.9 percentage points for school scenes and 1.2 points for home scenes, while shopping-mall results barely change. A GPS-level environment prior (school, home, or mall) prunes the candidate scene list in advance, which is why the method can run in real time. If the claim holds, robots in known buildings can tell rooms apart more reliably without retraining a bigger classifier.

What carries the argument

The carrying mechanism is a per-candidate context-vector scoring step, the procedure at the heart of Algorithm 1. For each candidate scene $s_j$ among the CNN's top-5 predictions, the image is represented as a weighted sum of the detected object vectors, $I_k(j) = \sum_i w_{i,j} o_i$; the weight matrix $W$ and the vocabulary vectors $V$ are themselves learned with a hinge-loss cosine-similarity training model, initialized from word2vec embeddings and tf-idf statistics. The refined ranking is produced by multiplying the CNN confidence $ACC_k$ by the normalized cosine similarity between $I_k(j)$ and the scene vector $s_j$, then sorting descending. The taxonomy is the second supporting mechanism: it prunes the Places365 label set to a small per-environment list, turning a 365-way classification problem into a 14-to-31-way problem for the chosen environment.

What would settle it

Run the same ResNet plus word-embedding re-ranker on the authors' test images while withholding the environment taxonomy, so every image is classified against all 365 Places classes or against the wrong environment's short list. If top-1 accuracy stays near 92%, the taxonomy is not load-bearing; if it collapses, the reported result is a pruned-search effect rather than a general scene-classification gain.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that the context of a scene—the objects a scene-parsing network finds in the image—carries signal that the scene classifier's top-5 confidence scores leave unused. The method represents each detected object and each candidate scene label as word vectors, forms an image vector per candidate as a weighted sum of the object vectors, and multiplies the CNN's confidences by normalized cosine similarities to those scene vectors. After this re-ranking, top-1 accuracy on the authors' test set rises from 90.33% to 92.25% for school and from 92.03% to 93.27% for home, while shopping mall stays essentially flat at 87.39%. The authors attribute the flat mall result to the scene-parsing model's failure to detect distinguishing objects such as shoes, watches, and hats. The method presumes a taxonomy that tells the robot which broad environment it is in, so the full set of 365 Places classes is reduced to 24 school scenes, 14 home scenes, or 31 mall scenes before classification.

Load-bearing premise

The method assumes the robot already knows which broad environment it is in (school, home, or shopping mall), so the scene classifier only has to choose among a short pruned list; if that prior is wrong or absent, the reported accuracies do not apply.

Editorial extensions

If this is right

In a known environment, the taxonomy cuts the classification problem from 365 classes to 14–31 classes, which is what makes the system fast enough for a robot.
The word-embedding refinement can break near-ties in the CNN's top-5 confidence scores, as in the chemistry-lab example where the CNN split its confidence across physics lab, chemistry lab, and gymnasium.
The refinement does not help when the scene parser cannot see the objects that separate the candidate scenes, as happened in the shopping-mall environment.
The same pipeline, with a per-environment ResNet and the context re-ranker, runs on a Pioneer3AT robot with a ZED camera and a laptop GPU.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Beyond the paper, the same re-ranking step could apply to any top-k classifier whose candidate labels can be paired with an object vocabulary, for example video scene tagging or indoor navigation landmark recognition.
The paper does not test transfer, but because the object and scene vectors are learned from a shared embedding space, the learned weights might transfer to new indoor scene labels without retraining the CNN.
A soft alternative the paper leaves untested is to replace the hard GPS environment prior with a probability distribution over environments, which would reveal how gracefully accuracy degrades when the robot is unsure where it is.
The flat shopping-mall result points to a concrete bottleneck the paper names but does not fix: the scene parser misses small distinguishing objects such as shoes and hats, so improving small-object detection is the direct route to gains there.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The paper combines a per-environment ResNet scene classifier with learned word2vec object embeddings to re-rank top-5 predictions. The specific synthesis is new; Object Bank used object detectors, not cosine similarity with a fine-tuned weight matrix. The algorithm is described clearly, and the robot deployment is a nice touch. They also sensibly merge and prune Places365 classes to deal with label noise.

The main problem is the evaluation. The claim that word2vec refinement improves accuracy rests on aggregate top-1 numbers from a test set of 10,000 frames captured as GoPro sequences from 69 places. Adjacent frames are near-duplicates, so the effective sample size is far smaller than 10,000. There are no per-place accuracies, no confidence intervals, and no significance tests. The school gain (90.33% to 92.25%) and home gain (92.03% to 93.27%) are about 1-2 points, which could easily be driven by a few places if some sequences happen to favor the re-ranking. Shopping mall ties the baseline (87.31% vs 87.39%), and the authors attribute this to the scene parser missing distinguishing objects. That is plausible, but it also shows the claimed benefit is inconsistent.

The stress-test concern about autocorrelation is on target. Without a per-place or per-sequence breakdown, I would not trust that the gain is real. The method itself is internally coherent; training is standard supervised learning, and there is no obvious circularity. But the central evidence is thin.

This paper is for researchers interested in lightweight refinements to scene classifiers for robots in known environments. It is not a breakthrough, but the idea is reasonable and the deployment is honest. A serious referee could ask for per-place error bars, leave-one-sequence-out validation, and release of the test set.

I would not desk-reject this. It deserves referee time, but acceptance should be conditional on a statistically grounded evaluation.

Referee Report

3 major / 6 minor

Summary. The paper proposes a two-stage indoor scene classification pipeline for robots. A taxonomy prunes the 365 Places classes down to a per-environment list (e.g., school, home, shopping mall). A ResNet50 trained on the reduced Places365 subset produces top-5 predictions; a pre-trained scene parsing model detects objects; word2vec embeddings are used to compute a weighted object vector for the image and cosine similarity to each predicted scene label, re-ranking the CNN scores. The weight matrix W and the refined embeddings are learned on the reduced Places365 training set. The authors evaluate on a self-built dataset of 10,000 images from 69 places (plus YouTube frames) and on a Places365 validation subset, reporting top-1 accuracy in Table II. They also describe a deployment on a Pioneer3AT robot.

Significance. If the reported gains are real, the paper makes a modest but useful contribution: fusing object-level context via word embeddings improves top-1 accuracy by 1-2 percentage points on the authors' test set for school and home scenes, while leaving shopping mall essentially unchanged, and the method is simple and compatible with a real-time robot pipeline. The ResNet50 baseline under the same reduced taxonomy is a fair comparison, and the algorithm is described clearly enough to reimplement. The strengths are the explicit algorithm, the use of standard components, and the honest acknowledgment that the shopping mall refinement is not noticeable. However, the evidence is currently not strong enough: the test set is small in effective size and not released, and the Places365 validation results show negligible or negative gains. The paper would be substantially strengthened by per-place accuracy, statistical tests, and full training details.

major comments (3)

[Section V, Table II] The central claim that the word2vec refinement improves accuracy is not statistically supported. The test set consists of 10,000 images from 69 places, captured as GoPro sequences (Section IV), so adjacent frames are near-duplicates and the effective sample size is far smaller than 10,000. The reported gains are 1.92 and 1.24 percentage points for school and home, while shopping mall is essentially unchanged (87.31% to 87.39%). On the Places365 validation subset the gains are only 0.46, 0.21, and -0.03 percentage points. No confidence intervals, per-place accuracy, or significance tests are given. The authors should report per-place and per-video-sequence numbers, a paired statistical test (e.g., McNemar or paired bootstrap), and a discussion of why the effect appears only on the self-built test set.
[Section III-C] The training of the weight matrix W and the refined vectors V is underspecified. Equation (1) defines the image vector as a weighted sum of object vectors, but the text does not state the optimization objective beyond "hinge loss with margin equal to 0.1", nor the learning rate, number of epochs, embedding dimension, or whether W and V are learned per environment or globally. It also does not specify whether the CNN predictions used to train W come from the same ResNet50 models evaluated in Table II, nor whether any validation data (including the self-built test set) is used to tune hyperparameters. Without these details, the reported accuracy gain cannot be reproduced or checked for possible leakage.
[Section I and Section III-A] The method assumes the environment type is known before classification, using GPS to prune the taxonomy. The paper never evaluates robustness to an incorrect or absent prior, and the conclusion states that robots can recognize places "with high confidence/accuracy" without this qualifier. The authors should either restrict their claims to the known-environment setting or report experiments with a mis-specified or open-set prior.

minor comments (6)

[Throughout] There are several typographical errors: "dinning room" in Section I, "representaion" in Section II, "predeﬁned paces" in Section II, and inconsistent use of "Place365" vs "Places365".
[Figure 4] In the caption, "Column AxB" should be typeset as "Column A × B" or "Column A*B" to avoid confusion.
[Section III-C] The notation "W = {w1,1, ..wi,j, ..w150,n}" has a typographical issue in the indexing display, and "IV ector k(j) = < 0 > vector" would be clearer as "IV ector k(j) = 0" with a zero vector.
[Section IV] The description of the YouTube images is vague: no sources, licenses, or frame selection procedure are given, which affects the reproducibility of the test set.
[Section V, Table II] Table II would benefit from reporting the number of test images per environment and per class, since class imbalance can distort aggregate top-1 accuracy.
[Section III-A] The sentence "The code and model parameters can be downloaded from this link" points to the Places365 website rather than a code repository; please correct or clarify the link.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the word2vec refinement is a supervised reranker trained on separate data, and the reported accuracy gain is not forced by construction.

full rationale

The paper's central claim is an empirical comparison between a ResNet50 baseline and a ResNet50-plus-word2vec reranker, and the derivation chain does not reduce to its own inputs. The object and scene vectors V and the weight matrix W are learned on the reduced Places365 training set (Section III-C: 'We pre-compute the training predictions, T, ACC and objects in images, IO. We feed this information into our cosine similarity training model ... to learn the weights, W and Vocabulary, V'), using a hinge loss, and are then evaluated on a separately captured GoPro/YouTube test set described in Section IV. The reranking rule RT_k = descendingOrder(ACC_k * Similarity) in Algorithm 1 combines CNN confidences with a cosine similarity computed from learned vectors, but nothing in that equation forces the final top-1 label to match the ground truth; the method can and does fail on some examples (Figure 4, rows 5-7), and the shopping-mall accuracy is essentially unchanged (87.31% vs 87.39%, Table II), confirming the comparison is not tautological. The only self-citations ([4], [20]) appear in related-work context and are not load-bearing for the method or the evaluation. Concerns about autocorrelated test sequences, small effective sample size, and missing error bars are statistical-validity concerns rather than circularity, so they do not raise the circularity score.

Assumptions & free parameters 3 free parameters · 3 assumptions · 0 invented entities

The central claim depends on the known-environment prior and on the scene parser being able to deliver discriminative object cues. Hyperparameters for the word2vec training (margin, learning rate, epochs, embedding size) are underreported. No new physical or conceptual entities are introduced.

free parameters (3)

hinge loss margin = 0.1
Margin used in the cosine similarity training model (Section III-C). Chosen by hand without sensitivity analysis.
word embedding dimension = not stated
Inherited from the pre-trained word2vec model (Gensim); not reported in the paper, so the reader cannot reproduce the vector space size.
word2vec fine-tuning learning rate and epochs = not stated
The paper says the vectors and weights are learned via cosine similarity training with hinge loss (Section III-C) but does not report the optimizer, learning rate, or number of iterations.

assumptions (3)

domain assumption The robot knows its broad environment (school, home, shopping mall) before classification, enabling taxonomy pruning.
Section I states the robot uses coarse GPS to reduce the search space; Section III builds per-environment models. If this prior is missing, the method must handle all 156 scenes and the reported numbers do not apply.
domain assumption The ADE20K scene parser's 150 object categories can detect the objects needed to discriminate scenes within an environment.
Section V notes the parser could not detect shoes, watches, and hats in shopping malls, and refinement for that environment is unchanged, showing the claim depends on parser object coverage.
domain assumption The reduced, merged version of Places365 remains a valid training distribution for real-world indoor images captured with a GoPro.
Section IV describes merging confusing classes and removing outdoor scenes; the resulting 156-class task is easier than the original 365-class task, and the self-built test set was labeled by the authors.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Scene Classification in Indoor Environments for Robots using Context Based Word Embeddings." pith.science (2026). https://pith.science/paper/GFS3FKAR

@misc{pith2026190806422,
  author       = {Pith},
  title        = {Pith review of: Scene Classification in Indoor Environments for Robots using Context Based Word Embeddings},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GFS3FKAR}},
  note         = {Machine review of arXiv:1908.06422}
}

read the original abstract

Scene Classification has been addressed with numerous techniques in computer vision literature. However, with the increasing number of scene classes in datasets in the field, it has become difficult to achieve high accuracy in the context of robotics. In this paper, we implement an approach which combines traditional deep learning techniques with natural language processing methods to generate a word embedding based Scene Classification algorithm. We use the key idea that context (objects in the scene) of an image should be representative of the scene label meaning a group of objects could assist to predict the scene class. Objects present in the scene are represented by vectors and the images are re-classified based on the objects present in the scene to refine the initial classification by a Convolutional Neural Network (CNN). In our approach we address indoor Scene Classification task using a model trained with a reduced pre-processed version of the Places365 dataset and an empirical analysis is done on a real-world dataset that we built by capturing image sequences using a GoPro camera. We also report results obtained on a subset of the Places365 dataset using our approach and additionally show a deployment of our approach on a robot operating in a real-world environment.

Figures

Figures reproduced from arXiv: 1908.06422 by the authors.

**Figure 2.** Overview of our algorithm. Our approach takes an RGB image as input. Then passes it onto two CNN modules. The top module is for Scene [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

**Figure 3.** Training Model for learning the new word2vector [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: Column A is the top-5 confidences obtained from the CNN Scene Classification module. Column B is the cosine similarity score from the Word [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

26 extracted references · 15 canonical work pages

[1]

Places: A 10 million image database for scene recognition,

B. Zhou, A. Lapedriza, A. Khosla, A. Oliva, and A. Torralba, “Places: A 10 million image database for scene recognition,” IEEE Transac- tions on Pattern Analysis and Machine Intelligence , 2017

2017
[2]

Sun database: Large-scale scene recognition from abbey to zoo,

J. Xiao, J. Hays, K. A. Ehinger, A. Oliva, and A. Torralba, “Sun database: Large-scale scene recognition from abbey to zoo,” in Com- puter vision and pattern recognition (CVPR), 2010 IEEE conference on. IEEE, 2010, pp. 3485–3492

work page 2010
[3]

Deep learning based place recognition for challenging environments,

D. Kumar, “Deep learning based place recognition for challenging environments,” Master’s thesis, University of Waterloo, 2016

work page 2016
[4]

Indoor place recognition system for localization of mobile robots,

R. Sahdev and J. K. Tsotsos, “Indoor place recognition system for localization of mobile robots,” in Computer and Robot Vision (CRV), 2016 13th Conference on . IEEE, 2016, pp. 53–60

work page 2016
[5]

Beyond bags of features: Spatial pyramid matching for recognizing natural scene categories,

S. Lazebnik, C. Schmid, and J. Ponce, “Beyond bags of features: Spatial pyramid matching for recognizing natural scene categories,” in Computer vision and pattern recognition, 2006 IEEE computer society conference on, vol. 2. IEEE, 2006, pp. 2169–2178

work page 2006
[6]

Efﬁcient estimation of word representations in vector space,

T. Mikolov, K. Chen, G. Corrado, and J. Dean, “Efﬁcient estimation of word representations in vector space,”arXiv preprint arXiv:1301.3781, 2013

arXiv 2013
[7]

Object bank: A high- level image representation for scene classiﬁcation & semantic feature sparsiﬁcation,

L.-J. Li, H. Su, L. Fei-Fei, and E. P. Xing, “Object bank: A high- level image representation for scene classiﬁcation & semantic feature sparsiﬁcation,” in Advances in neural information processing systems , 2010, pp. 1378–1386

work page 2010
[8]

Evaluating bag-of-visual-words representations in scene classiﬁcation,

J. Yang, Y .-G. Jiang, A. G. Hauptmann, and C.-W. Ngo, “Evaluating bag-of-visual-words representations in scene classiﬁcation,” in Pro- ceedings of the international workshop on Workshop on multimedia information retrieval. ACM, 2007, pp. 197–206

work page 2007

Show all 26 references

[9]

Scene classiﬁcation using a hybrid generative/discriminative approach,

A. Bosch, A. Zisserman, and X. Mu ˜noz, “Scene classiﬁcation using a hybrid generative/discriminative approach,” IEEE transactions on pattern analysis and machine intelligence , vol. 30, no. 4, pp. 712– 727, 2008

2008
[10]

Imagenet classiﬁcation with deep convolutional neural networks,

A. Krizhevsky, I. Sutskever, and G. E. Hinton, “Imagenet classiﬁcation with deep convolutional neural networks,” in Advances in neural information processing systems , 2012, pp. 1097–1105

2012
[11]

Very deep convolutional networks for large-scale image recognition,

K. Simonyan and A. Zisserman, “Very deep convolutional networks for large-scale image recognition,” arXiv preprint arXiv:1409.1556 , 2014

2014 arXiv
[12]

Going deeper with convolutions,

C. Szegedy, W. Liu, Y . Jia, P. Sermanet, S. Reed, D. Anguelov, D. Erhan, V . Vanhoucke, and A. Rabinovich, “Going deeper with convolutions,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2015, pp. 1–9

2015
[13]

Scene categorization with spec- tral features,

S. H. Khan, M. Hayat, and F. Porikli, “Scene categorization with spec- tral features,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , 2017, pp. 5638–5648

2017
[14]

Segnet: A deep convolutional encoder-decoder architecture for image segmentation,

V . Badrinarayanan, A. Kendall, and R. Cipolla, “Segnet: A deep convolutional encoder-decoder architecture for image segmentation,” IEEE Transactions on Pattern Analysis and Machine Intelligence , 2017

2017
[15]

Scene parsing through ade20k dataset,

B. Zhou, H. Zhao, X. Puig, S. Fidler, A. Barriuso, and A. Torralba, “Scene parsing through ade20k dataset,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , 2017

2017
[16]

Yolo9000: better, faster, stronger,

J. Redmon and A. Farhadi, “Yolo9000: better, faster, stronger,” arXiv preprint arXiv:1612.08242, 2016

2016 arXiv
[17]

Glove: Global vectors for word representation,

J. Pennington, R. Socher, and C. Manning, “Glove: Global vectors for word representation,” in Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP) , 2014, pp. 1532–1543

2014
[18]

Bidirectional lstm-crf models for sequence tagging,

Z. Huang, W. Xu, and K. Yu, “Bidirectional lstm-crf models for sequence tagging,” arXiv preprint arXiv:1508.01991 , 2015

2015 arXiv
[19]

Neural architectures for named entity recognition,

G. Lample, M. Ballesteros, S. Subramanian, K. Kawakami, and C. Dyer, “Neural architectures for named entity recognition,” arXiv preprint arXiv:1603.01360, 2016. Fig. 4. Column A is the top-5 conﬁdences obtained from the CNN Scene Classiﬁcation module. Column B is the cosine si...

2016 arXiv
[20]

Detecting the magnitude of events from news articles,

A. Agrawal, R. Sahdev, H. Davoudi, F. Khonsari, A. An, and S. Mc- Grath, “Detecting the magnitude of events from news articles,” in Web Intelligence (WI), 2016 IEEE/WIC/ACM International Conference on . IEEE, 2016, pp. 177–184

2016
[21]

A local detection approach for named entity recognition and mention detection,

M. Xu, H. Jiang, and S. Watcharawittayakul, “A local detection approach for named entity recognition and mention detection,” in Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), vol. 1, 2017, pp. 1237–1247

2017
[22]

Bocnf: efﬁcient image matching with bag of convnet features for scalable and robust visual place recognition,

Y . Hou, H. Zhang, and S. Zhou, “Bocnf: efﬁcient image matching with bag of convnet features for scalable and robust visual place recognition,” Autonomous Robots, pp. 1–17, 2017

2017
[23]

Illumination invariant representation of natural images for visual place recognition,

M. Shakeri and H. Zhang, “Illumination invariant representation of natural images for visual place recognition,” in Intelligent Robots and Systems (IROS), 2016 IEEE/RSJ International Conference on . IEEE, 2016, pp. 466–472

2016
[24]

Learning deep features for scene recognition using places database,

B. Zhou, A. Lapedriza, J. Xiao, A. Torralba, and A. Oliva, “Learning deep features for scene recognition using places database,” inAdvances in neural information processing systems , 2014, pp. 487–495

2014
[25]

The pascal visual object classes (voc) challenge,

M. Everingham, L. Van Gool, C. K. Williams, J. Winn, and A. Zisser- man, “The pascal visual object classes (voc) challenge,” International journal of computer vision , vol. 88, no. 2, pp. 303–338, 2010

2010
[26]

Microsoft coco: Common objects in context,

T.-Y . Lin, M. Maire, S. Belongie, J. Hays, P. Perona, D. Ramanan, P. Doll ´ar, and C. L. Zitnick, “Microsoft coco: Common objects in context,” in European conference on computer vision. Springer, 2014, pp. 740–755

2014

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

Places: A 10 million image database for scene recognition,

B. Zhou, A. Lapedriza, A. Khosla, A. Oliva, and A. Torralba, “Places: A 10 million image database for scene recognition,” IEEE Transac- tions on Pattern Analysis and Machine Intelligence , 2017

2017

[2] [2]

Sun database: Large-scale scene recognition from abbey to zoo,

J. Xiao, J. Hays, K. A. Ehinger, A. Oliva, and A. Torralba, “Sun database: Large-scale scene recognition from abbey to zoo,” in Com- puter vision and pattern recognition (CVPR), 2010 IEEE conference on. IEEE, 2010, pp. 3485–3492

work page 2010

[3] [3]

Deep learning based place recognition for challenging environments,

D. Kumar, “Deep learning based place recognition for challenging environments,” Master’s thesis, University of Waterloo, 2016

work page 2016

[4] [4]

Indoor place recognition system for localization of mobile robots,

R. Sahdev and J. K. Tsotsos, “Indoor place recognition system for localization of mobile robots,” in Computer and Robot Vision (CRV), 2016 13th Conference on . IEEE, 2016, pp. 53–60

work page 2016

[5] [5]

Beyond bags of features: Spatial pyramid matching for recognizing natural scene categories,

S. Lazebnik, C. Schmid, and J. Ponce, “Beyond bags of features: Spatial pyramid matching for recognizing natural scene categories,” in Computer vision and pattern recognition, 2006 IEEE computer society conference on, vol. 2. IEEE, 2006, pp. 2169–2178

work page 2006

[6] [6]

Efﬁcient estimation of word representations in vector space,

T. Mikolov, K. Chen, G. Corrado, and J. Dean, “Efﬁcient estimation of word representations in vector space,”arXiv preprint arXiv:1301.3781, 2013

arXiv 2013

[7] [7]

Object bank: A high- level image representation for scene classiﬁcation & semantic feature sparsiﬁcation,

L.-J. Li, H. Su, L. Fei-Fei, and E. P. Xing, “Object bank: A high- level image representation for scene classiﬁcation & semantic feature sparsiﬁcation,” in Advances in neural information processing systems , 2010, pp. 1378–1386

work page 2010

[8] [8]

Evaluating bag-of-visual-words representations in scene classiﬁcation,

J. Yang, Y .-G. Jiang, A. G. Hauptmann, and C.-W. Ngo, “Evaluating bag-of-visual-words representations in scene classiﬁcation,” in Pro- ceedings of the international workshop on Workshop on multimedia information retrieval. ACM, 2007, pp. 197–206

work page 2007